polars-df 0.13.0-x64-mingw-ucrt
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.yardopts +3 -0
- data/CHANGELOG.md +208 -0
- data/Cargo.lock +2556 -0
- data/Cargo.toml +6 -0
- data/LICENSE-THIRD-PARTY.txt +39278 -0
- data/LICENSE.txt +20 -0
- data/README.md +437 -0
- data/lib/polars/3.1/polars.so +0 -0
- data/lib/polars/3.2/polars.so +0 -0
- data/lib/polars/3.3/polars.so +0 -0
- data/lib/polars/array_expr.rb +537 -0
- data/lib/polars/array_name_space.rb +423 -0
- data/lib/polars/batched_csv_reader.rb +104 -0
- data/lib/polars/binary_expr.rb +77 -0
- data/lib/polars/binary_name_space.rb +66 -0
- data/lib/polars/cat_expr.rb +36 -0
- data/lib/polars/cat_name_space.rb +88 -0
- data/lib/polars/config.rb +530 -0
- data/lib/polars/convert.rb +98 -0
- data/lib/polars/data_frame.rb +5191 -0
- data/lib/polars/data_types.rb +466 -0
- data/lib/polars/date_time_expr.rb +1397 -0
- data/lib/polars/date_time_name_space.rb +1287 -0
- data/lib/polars/dynamic_group_by.rb +52 -0
- data/lib/polars/exceptions.rb +38 -0
- data/lib/polars/expr.rb +7256 -0
- data/lib/polars/expr_dispatch.rb +22 -0
- data/lib/polars/functions/aggregation/horizontal.rb +246 -0
- data/lib/polars/functions/aggregation/vertical.rb +282 -0
- data/lib/polars/functions/as_datatype.rb +271 -0
- data/lib/polars/functions/col.rb +47 -0
- data/lib/polars/functions/eager.rb +182 -0
- data/lib/polars/functions/lazy.rb +1329 -0
- data/lib/polars/functions/len.rb +49 -0
- data/lib/polars/functions/lit.rb +35 -0
- data/lib/polars/functions/random.rb +16 -0
- data/lib/polars/functions/range/date_range.rb +136 -0
- data/lib/polars/functions/range/datetime_range.rb +149 -0
- data/lib/polars/functions/range/int_range.rb +51 -0
- data/lib/polars/functions/range/time_range.rb +141 -0
- data/lib/polars/functions/repeat.rb +144 -0
- data/lib/polars/functions/whenthen.rb +96 -0
- data/lib/polars/functions.rb +57 -0
- data/lib/polars/group_by.rb +613 -0
- data/lib/polars/io/avro.rb +24 -0
- data/lib/polars/io/csv.rb +696 -0
- data/lib/polars/io/database.rb +73 -0
- data/lib/polars/io/ipc.rb +275 -0
- data/lib/polars/io/json.rb +29 -0
- data/lib/polars/io/ndjson.rb +80 -0
- data/lib/polars/io/parquet.rb +233 -0
- data/lib/polars/lazy_frame.rb +2708 -0
- data/lib/polars/lazy_group_by.rb +181 -0
- data/lib/polars/list_expr.rb +791 -0
- data/lib/polars/list_name_space.rb +449 -0
- data/lib/polars/meta_expr.rb +222 -0
- data/lib/polars/name_expr.rb +198 -0
- data/lib/polars/plot.rb +109 -0
- data/lib/polars/rolling_group_by.rb +35 -0
- data/lib/polars/series.rb +4444 -0
- data/lib/polars/slice.rb +104 -0
- data/lib/polars/sql_context.rb +194 -0
- data/lib/polars/string_cache.rb +75 -0
- data/lib/polars/string_expr.rb +1495 -0
- data/lib/polars/string_name_space.rb +811 -0
- data/lib/polars/struct_expr.rb +98 -0
- data/lib/polars/struct_name_space.rb +96 -0
- data/lib/polars/testing.rb +507 -0
- data/lib/polars/utils/constants.rb +9 -0
- data/lib/polars/utils/convert.rb +97 -0
- data/lib/polars/utils/parse.rb +89 -0
- data/lib/polars/utils/various.rb +76 -0
- data/lib/polars/utils/wrap.rb +19 -0
- data/lib/polars/utils.rb +130 -0
- data/lib/polars/version.rb +4 -0
- data/lib/polars/whenthen.rb +83 -0
- data/lib/polars-df.rb +1 -0
- data/lib/polars.rb +91 -0
- metadata +138 -0
@@ -0,0 +1,4444 @@
|
|
1
|
+
module Polars
|
2
|
+
# A Series represents a single column in a polars DataFrame.
|
3
|
+
class Series
|
4
|
+
include ExprDispatch
|
5
|
+
|
6
|
+
# Create a new Series.
|
7
|
+
#
|
8
|
+
# @param name [String, Array, nil]
|
9
|
+
# Name of the series. Will be used as a column name when used in a DataFrame.
|
10
|
+
# When not specified, name is set to an empty string.
|
11
|
+
# @param values [Array, nil]
|
12
|
+
# One-dimensional data in various forms. Supported are: Array and Series.
|
13
|
+
# @param dtype [Symbol, nil]
|
14
|
+
# Polars dtype of the Series data. If not specified, the dtype is inferred.
|
15
|
+
# @param strict [Boolean]
|
16
|
+
# Throw error on numeric overflow.
|
17
|
+
# @param nan_to_null [Boolean]
|
18
|
+
# Not used.
|
19
|
+
# @param dtype_if_empty [Symbol, nil]
|
20
|
+
# If no dtype is specified and values contains `nil` or an empty array,
|
21
|
+
# set the Polars dtype of the Series data. If not specified, Float32 is used.
|
22
|
+
#
|
23
|
+
# @example Constructing a Series by specifying name and values positionally:
|
24
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
25
|
+
#
|
26
|
+
# @example Notice that the dtype is automatically inferred as a polars `Int64`:
|
27
|
+
# s.dtype
|
28
|
+
# # => Polars::Int64
|
29
|
+
#
|
30
|
+
# @example Constructing a Series with a specific dtype:
|
31
|
+
# s2 = Polars::Series.new("a", [1, 2, 3], dtype: :f32)
|
32
|
+
#
|
33
|
+
# @example It is possible to construct a Series with values as the first positional argument. This syntax considered an anti-pattern, but it can be useful in certain scenarios. You must specify any other arguments through keywords.
|
34
|
+
# s3 = Polars::Series.new([1, 2, 3])
|
35
|
+
def initialize(name = nil, values = nil, dtype: nil, strict: true, nan_to_null: false, dtype_if_empty: nil)
|
36
|
+
# Handle case where values are passed as the first argument
|
37
|
+
if !name.nil? && !name.is_a?(::String)
|
38
|
+
if values.nil?
|
39
|
+
values = name
|
40
|
+
name = nil
|
41
|
+
else
|
42
|
+
raise ArgumentError, "Series name must be a string."
|
43
|
+
end
|
44
|
+
end
|
45
|
+
|
46
|
+
name = "" if name.nil?
|
47
|
+
|
48
|
+
# TODO improve
|
49
|
+
if values.is_a?(Range) && values.begin.is_a?(::String)
|
50
|
+
values = values.to_a
|
51
|
+
end
|
52
|
+
|
53
|
+
if values.nil?
|
54
|
+
self._s = sequence_to_rbseries(name, [], dtype: dtype, dtype_if_empty: dtype_if_empty)
|
55
|
+
elsif values.is_a?(Series)
|
56
|
+
self._s = series_to_rbseries(name, values)
|
57
|
+
elsif values.is_a?(Range)
|
58
|
+
self._s =
|
59
|
+
Polars.arange(
|
60
|
+
values.first,
|
61
|
+
values.last + (values.exclude_end? ? 0 : 1),
|
62
|
+
step: 1,
|
63
|
+
eager: true,
|
64
|
+
dtype: dtype
|
65
|
+
)
|
66
|
+
.rename(name, in_place: true)
|
67
|
+
._s
|
68
|
+
elsif values.is_a?(::Array)
|
69
|
+
self._s = sequence_to_rbseries(name, values, dtype: dtype, strict: strict, dtype_if_empty: dtype_if_empty)
|
70
|
+
elsif defined?(Numo::NArray) && values.is_a?(Numo::NArray)
|
71
|
+
self._s = numo_to_rbseries(name, values, strict: strict, nan_to_null: nan_to_null)
|
72
|
+
|
73
|
+
if !dtype.nil?
|
74
|
+
self._s = self.cast(dtype, strict: true)._s
|
75
|
+
end
|
76
|
+
else
|
77
|
+
raise ArgumentError, "Series constructor called with unsupported type; got #{values.class.name}"
|
78
|
+
end
|
79
|
+
end
|
80
|
+
|
81
|
+
# @private
|
82
|
+
def self._from_rbseries(s)
|
83
|
+
series = Series.allocate
|
84
|
+
series._s = s
|
85
|
+
series
|
86
|
+
end
|
87
|
+
|
88
|
+
# Get the data type of this Series.
|
89
|
+
#
|
90
|
+
# @return [Symbol]
|
91
|
+
def dtype
|
92
|
+
_s.dtype
|
93
|
+
end
|
94
|
+
|
95
|
+
# Get flags that are set on the Series.
|
96
|
+
#
|
97
|
+
# @return [Hash]
|
98
|
+
def flags
|
99
|
+
out = {
|
100
|
+
"SORTED_ASC" => _s.is_sorted_flag,
|
101
|
+
"SORTED_DESC" => _s.is_sorted_reverse_flag
|
102
|
+
}
|
103
|
+
if dtype.is_a?(List)
|
104
|
+
out["FAST_EXPLODE"] = _s.can_fast_explode_flag
|
105
|
+
end
|
106
|
+
out
|
107
|
+
end
|
108
|
+
|
109
|
+
# Get the inner dtype in of a List typed Series.
|
110
|
+
#
|
111
|
+
# @return [Symbol]
|
112
|
+
def inner_dtype
|
113
|
+
_s.inner_dtype
|
114
|
+
end
|
115
|
+
|
116
|
+
# Get the name of this Series.
|
117
|
+
#
|
118
|
+
# @return [String]
|
119
|
+
def name
|
120
|
+
_s.name
|
121
|
+
end
|
122
|
+
|
123
|
+
# Shape of this Series.
|
124
|
+
#
|
125
|
+
# @return [Array]
|
126
|
+
def shape
|
127
|
+
[_s.len]
|
128
|
+
end
|
129
|
+
|
130
|
+
# Get the time unit of underlying Datetime Series as `"ns"`, `"us"`, or `"ms"`.
|
131
|
+
#
|
132
|
+
# @return [String]
|
133
|
+
def time_unit
|
134
|
+
_s.time_unit
|
135
|
+
end
|
136
|
+
|
137
|
+
# Returns a string representing the Series.
|
138
|
+
#
|
139
|
+
# @return [String]
|
140
|
+
def to_s
|
141
|
+
_s.to_s
|
142
|
+
end
|
143
|
+
alias_method :inspect, :to_s
|
144
|
+
|
145
|
+
# Bitwise AND.
|
146
|
+
#
|
147
|
+
# @return [Series]
|
148
|
+
def &(other)
|
149
|
+
if !other.is_a?(Series)
|
150
|
+
other = Series.new([other])
|
151
|
+
end
|
152
|
+
Utils.wrap_s(_s.bitand(other._s))
|
153
|
+
end
|
154
|
+
|
155
|
+
# Bitwise OR.
|
156
|
+
#
|
157
|
+
# @return [Series]
|
158
|
+
def |(other)
|
159
|
+
if !other.is_a?(Series)
|
160
|
+
other = Series.new([other])
|
161
|
+
end
|
162
|
+
Utils.wrap_s(_s.bitor(other._s))
|
163
|
+
end
|
164
|
+
|
165
|
+
# Bitwise XOR.
|
166
|
+
#
|
167
|
+
# @return [Series]
|
168
|
+
def ^(other)
|
169
|
+
if !other.is_a?(Series)
|
170
|
+
other = Series.new([other])
|
171
|
+
end
|
172
|
+
Utils.wrap_s(_s.bitxor(other._s))
|
173
|
+
end
|
174
|
+
|
175
|
+
# Equal.
|
176
|
+
#
|
177
|
+
# @return [Series]
|
178
|
+
def ==(other)
|
179
|
+
_comp(other, :eq)
|
180
|
+
end
|
181
|
+
|
182
|
+
# Not equal.
|
183
|
+
#
|
184
|
+
# @return [Series]
|
185
|
+
def !=(other)
|
186
|
+
_comp(other, :neq)
|
187
|
+
end
|
188
|
+
|
189
|
+
# Greater than.
|
190
|
+
#
|
191
|
+
# @return [Series]
|
192
|
+
def >(other)
|
193
|
+
_comp(other, :gt)
|
194
|
+
end
|
195
|
+
|
196
|
+
# Less than.
|
197
|
+
#
|
198
|
+
# @return [Series]
|
199
|
+
def <(other)
|
200
|
+
_comp(other, :lt)
|
201
|
+
end
|
202
|
+
|
203
|
+
# Greater than or equal.
|
204
|
+
#
|
205
|
+
# @return [Series]
|
206
|
+
def >=(other)
|
207
|
+
_comp(other, :gt_eq)
|
208
|
+
end
|
209
|
+
|
210
|
+
# Less than or equal.
|
211
|
+
#
|
212
|
+
# @return [Series]
|
213
|
+
def <=(other)
|
214
|
+
_comp(other, :lt_eq)
|
215
|
+
end
|
216
|
+
|
217
|
+
# Method equivalent of operator expression `series <= other`.
|
218
|
+
#
|
219
|
+
# @return [Series]
|
220
|
+
def le(other)
|
221
|
+
self <= other
|
222
|
+
end
|
223
|
+
|
224
|
+
# Method equivalent of operator expression `series < other`.
|
225
|
+
#
|
226
|
+
# @return [Series]
|
227
|
+
def lt(other)
|
228
|
+
self < other
|
229
|
+
end
|
230
|
+
|
231
|
+
# Method equivalent of operator expression `series == other`.
|
232
|
+
#
|
233
|
+
# @return [Series]
|
234
|
+
def eq(other)
|
235
|
+
self == other
|
236
|
+
end
|
237
|
+
|
238
|
+
# Method equivalent of equality operator `series == other` where `nil == nil`.
|
239
|
+
#
|
240
|
+
# This differs from the standard `ne` where null values are propagated.
|
241
|
+
#
|
242
|
+
# @param other [Object]
|
243
|
+
# A literal or expression value to compare with.
|
244
|
+
#
|
245
|
+
# @return [Object]
|
246
|
+
#
|
247
|
+
# @example
|
248
|
+
# s1 = Polars::Series.new("a", [333, 200, nil])
|
249
|
+
# s2 = Polars::Series.new("a", [100, 200, nil])
|
250
|
+
# s1.eq(s2)
|
251
|
+
# # =>
|
252
|
+
# # shape: (3,)
|
253
|
+
# # Series: 'a' [bool]
|
254
|
+
# # [
|
255
|
+
# # false
|
256
|
+
# # true
|
257
|
+
# # null
|
258
|
+
# # ]
|
259
|
+
#
|
260
|
+
# @example
|
261
|
+
# s1.eq_missing(s2)
|
262
|
+
# # =>
|
263
|
+
# # shape: (3,)
|
264
|
+
# # Series: 'a' [bool]
|
265
|
+
# # [
|
266
|
+
# # false
|
267
|
+
# # true
|
268
|
+
# # true
|
269
|
+
# # ]
|
270
|
+
def eq_missing(other)
|
271
|
+
if other.is_a?(Expr)
|
272
|
+
return Polars.lit(self).eq_missing(other)
|
273
|
+
end
|
274
|
+
to_frame.select(Polars.col(name).eq_missing(other)).to_series
|
275
|
+
end
|
276
|
+
|
277
|
+
# Method equivalent of operator expression `series != other`.
|
278
|
+
#
|
279
|
+
# @return [Series]
|
280
|
+
def ne(other)
|
281
|
+
self != other
|
282
|
+
end
|
283
|
+
|
284
|
+
# Method equivalent of equality operator `series != other` where `None == None`.
|
285
|
+
#
|
286
|
+
# This differs from the standard `ne` where null values are propagated.
|
287
|
+
#
|
288
|
+
# @param other [Object]
|
289
|
+
# A literal or expression value to compare with.
|
290
|
+
#
|
291
|
+
# @return [Object]
|
292
|
+
#
|
293
|
+
# @example
|
294
|
+
# s1 = Polars::Series.new("a", [333, 200, nil])
|
295
|
+
# s2 = Polars::Series.new("a", [100, 200, nil])
|
296
|
+
# s1.ne(s2)
|
297
|
+
# # =>
|
298
|
+
# # shape: (3,)
|
299
|
+
# # Series: 'a' [bool]
|
300
|
+
# # [
|
301
|
+
# # true
|
302
|
+
# # false
|
303
|
+
# # null
|
304
|
+
# # ]
|
305
|
+
#
|
306
|
+
# @example
|
307
|
+
# s1.ne_missing(s2)
|
308
|
+
# # =>
|
309
|
+
# # shape: (3,)
|
310
|
+
# # Series: 'a' [bool]
|
311
|
+
# # [
|
312
|
+
# # true
|
313
|
+
# # false
|
314
|
+
# # false
|
315
|
+
# # ]
|
316
|
+
def ne_missing(other)
|
317
|
+
if other.is_a?(Expr)
|
318
|
+
return Polars.lit(self).ne_missing(other)
|
319
|
+
end
|
320
|
+
to_frame.select(Polars.col(name).ne_missing(other)).to_series
|
321
|
+
end
|
322
|
+
|
323
|
+
# Method equivalent of operator expression `series >= other`.
|
324
|
+
#
|
325
|
+
# @return [Series]
|
326
|
+
def ge(other)
|
327
|
+
self >= other
|
328
|
+
end
|
329
|
+
|
330
|
+
# Method equivalent of operator expression `series > other`.
|
331
|
+
#
|
332
|
+
# @return [Series]
|
333
|
+
def gt(other)
|
334
|
+
self > other
|
335
|
+
end
|
336
|
+
|
337
|
+
# Performs addition.
|
338
|
+
#
|
339
|
+
# @return [Series]
|
340
|
+
def +(other)
|
341
|
+
_arithmetic(other, :add)
|
342
|
+
end
|
343
|
+
|
344
|
+
# Performs subtraction.
|
345
|
+
#
|
346
|
+
# @return [Series]
|
347
|
+
def -(other)
|
348
|
+
_arithmetic(other, :sub)
|
349
|
+
end
|
350
|
+
|
351
|
+
# Performs multiplication.
|
352
|
+
#
|
353
|
+
# @return [Series]
|
354
|
+
def *(other)
|
355
|
+
if is_temporal
|
356
|
+
raise ArgumentError, "first cast to integer before multiplying datelike dtypes"
|
357
|
+
elsif other.is_a?(DataFrame)
|
358
|
+
other * self
|
359
|
+
else
|
360
|
+
_arithmetic(other, :mul)
|
361
|
+
end
|
362
|
+
end
|
363
|
+
|
364
|
+
# Performs division.
|
365
|
+
#
|
366
|
+
# @return [Series]
|
367
|
+
def /(other)
|
368
|
+
if is_temporal
|
369
|
+
raise ArgumentError, "first cast to integer before dividing datelike dtypes"
|
370
|
+
end
|
371
|
+
|
372
|
+
if is_float
|
373
|
+
return _arithmetic(other, :div)
|
374
|
+
end
|
375
|
+
|
376
|
+
cast(Float64) / other
|
377
|
+
end
|
378
|
+
|
379
|
+
# Returns the modulo.
|
380
|
+
#
|
381
|
+
# @return [Series]
|
382
|
+
def %(other)
|
383
|
+
if is_datelike
|
384
|
+
raise ArgumentError, "first cast to integer before applying modulo on datelike dtypes"
|
385
|
+
end
|
386
|
+
_arithmetic(other, :rem)
|
387
|
+
end
|
388
|
+
|
389
|
+
# Raises to the power of exponent.
|
390
|
+
#
|
391
|
+
# @return [Series]
|
392
|
+
def **(power)
|
393
|
+
if is_datelike
|
394
|
+
raise ArgumentError, "first cast to integer before raising datelike dtypes to a power"
|
395
|
+
end
|
396
|
+
to_frame.select(Polars.col(name).pow(power)).to_series
|
397
|
+
end
|
398
|
+
|
399
|
+
# Performs boolean not.
|
400
|
+
#
|
401
|
+
# @return [Series]
|
402
|
+
def !
|
403
|
+
if dtype == Boolean
|
404
|
+
return Utils.wrap_s(_s.not)
|
405
|
+
end
|
406
|
+
raise NotImplementedError
|
407
|
+
end
|
408
|
+
|
409
|
+
# Performs negation.
|
410
|
+
#
|
411
|
+
# @return [Series]
|
412
|
+
def -@
|
413
|
+
0 - self
|
414
|
+
end
|
415
|
+
|
416
|
+
# Returns an enumerator.
|
417
|
+
#
|
418
|
+
# @return [Object]
|
419
|
+
def each
|
420
|
+
return to_enum(:each) unless block_given?
|
421
|
+
|
422
|
+
length.times do |i|
|
423
|
+
yield self[i]
|
424
|
+
end
|
425
|
+
end
|
426
|
+
|
427
|
+
# Returns elements of the Series.
|
428
|
+
#
|
429
|
+
# @return [Object]
|
430
|
+
def [](item)
|
431
|
+
if item.is_a?(Series) && [UInt8, UInt16, UInt32, UInt64, Int8, Int16, Int32, Int64].include?(item.dtype)
|
432
|
+
return Utils.wrap_s(_s.take_with_series(_pos_idxs(item)._s))
|
433
|
+
end
|
434
|
+
|
435
|
+
if item.is_a?(Series) && item.bool?
|
436
|
+
return filter(item)
|
437
|
+
end
|
438
|
+
|
439
|
+
if item.is_a?(Integer)
|
440
|
+
if item < 0
|
441
|
+
item = len + item
|
442
|
+
end
|
443
|
+
|
444
|
+
return _s.get_idx(item)
|
445
|
+
end
|
446
|
+
|
447
|
+
if item.is_a?(Range)
|
448
|
+
return Slice.new(self).apply(item)
|
449
|
+
end
|
450
|
+
|
451
|
+
if Utils.is_int_sequence(item)
|
452
|
+
return Utils.wrap_s(_s.take_with_series(_pos_idxs(Series.new("", item))._s))
|
453
|
+
end
|
454
|
+
|
455
|
+
raise ArgumentError, "Cannot get item of type: #{item.class.name}"
|
456
|
+
end
|
457
|
+
|
458
|
+
# Sets an element of the Series.
|
459
|
+
#
|
460
|
+
# @return [Object]
|
461
|
+
def []=(key, value)
|
462
|
+
if value.is_a?(::Array)
|
463
|
+
if is_numeric || is_datelike
|
464
|
+
scatter(key, value)
|
465
|
+
return
|
466
|
+
end
|
467
|
+
raise ArgumentError, "cannot set Series of dtype: #{dtype} with list/tuple as value; use a scalar value"
|
468
|
+
end
|
469
|
+
|
470
|
+
if key.is_a?(Series)
|
471
|
+
if key.dtype == Boolean
|
472
|
+
self._s = set(key, value)._s
|
473
|
+
elsif key.dtype == UInt64
|
474
|
+
self._s = scatter(key.cast(UInt32), value)._s
|
475
|
+
elsif key.dtype == UInt32
|
476
|
+
self._s = scatter(key, value)._s
|
477
|
+
else
|
478
|
+
raise Todo
|
479
|
+
end
|
480
|
+
elsif key.is_a?(::Array)
|
481
|
+
s = Utils.wrap_s(sequence_to_rbseries("", key, dtype: UInt32))
|
482
|
+
self[s] = value
|
483
|
+
elsif key.is_a?(Range)
|
484
|
+
s = Series.new("", key, dtype: UInt32)
|
485
|
+
self[s] = value
|
486
|
+
elsif key.is_a?(Integer)
|
487
|
+
self[[key]] = value
|
488
|
+
else
|
489
|
+
raise ArgumentError, "cannot use #{key} for indexing"
|
490
|
+
end
|
491
|
+
end
|
492
|
+
|
493
|
+
# Return an estimation of the total (heap) allocated size of the Series.
|
494
|
+
#
|
495
|
+
# Estimated size is given in the specified unit (bytes by default).
|
496
|
+
#
|
497
|
+
# This estimation is the sum of the size of its buffers, validity, including
|
498
|
+
# nested arrays. Multiple arrays may share buffers and bitmaps. Therefore, the
|
499
|
+
# size of 2 arrays is not the sum of the sizes computed from this function. In
|
500
|
+
# particular, StructArray's size is an upper bound.
|
501
|
+
#
|
502
|
+
# When an array is sliced, its allocated size remains constant because the buffer
|
503
|
+
# unchanged. However, this function will yield a smaller number. This is because
|
504
|
+
# this function returns the visible size of the buffer, not its total capacity.
|
505
|
+
#
|
506
|
+
# FFI buffers are included in this estimation.
|
507
|
+
#
|
508
|
+
# @param unit ["b", "kb", "mb", "gb", "tb"]
|
509
|
+
# Scale the returned size to the given unit.
|
510
|
+
#
|
511
|
+
# @return [Numeric]
|
512
|
+
#
|
513
|
+
# @example
|
514
|
+
# s = Polars::Series.new("values", 1..1_000_000, dtype: :u32)
|
515
|
+
# s.estimated_size
|
516
|
+
# # => 4000000
|
517
|
+
# s.estimated_size("mb")
|
518
|
+
# # => 3.814697265625
|
519
|
+
def estimated_size(unit = "b")
|
520
|
+
sz = _s.estimated_size
|
521
|
+
Utils.scale_bytes(sz, to: unit)
|
522
|
+
end
|
523
|
+
|
524
|
+
# Compute the square root of the elements.
|
525
|
+
#
|
526
|
+
# @return [Series]
|
527
|
+
def sqrt
|
528
|
+
self**0.5
|
529
|
+
end
|
530
|
+
|
531
|
+
# Check if any boolean value in the column is `true`.
|
532
|
+
#
|
533
|
+
# @return [Boolean]
|
534
|
+
def any?(ignore_nulls: true, &block)
|
535
|
+
if block_given?
|
536
|
+
apply(skip_nulls: ignore_nulls, &block).any?
|
537
|
+
else
|
538
|
+
_s.any(ignore_nulls)
|
539
|
+
end
|
540
|
+
end
|
541
|
+
alias_method :any, :any?
|
542
|
+
|
543
|
+
# Check if all boolean values in the column are `true`.
|
544
|
+
#
|
545
|
+
# @return [Boolean]
|
546
|
+
def all?(ignore_nulls: true, &block)
|
547
|
+
if block_given?
|
548
|
+
apply(skip_nulls: ignore_nulls, &block).all?
|
549
|
+
else
|
550
|
+
_s.all(ignore_nulls)
|
551
|
+
end
|
552
|
+
end
|
553
|
+
alias_method :all, :all?
|
554
|
+
|
555
|
+
# Check if all boolean values in the column are `false`.
|
556
|
+
#
|
557
|
+
# @return [Boolean]
|
558
|
+
def none?(&block)
|
559
|
+
if block_given?
|
560
|
+
apply(&block).none?
|
561
|
+
else
|
562
|
+
to_frame.select(Polars.col(name).is_not.all).to_series[0]
|
563
|
+
end
|
564
|
+
end
|
565
|
+
alias_method :none, :none?
|
566
|
+
|
567
|
+
# Compute the logarithm to a given base.
|
568
|
+
#
|
569
|
+
# @param base [Float]
|
570
|
+
# Given base, defaults to `Math::E`.
|
571
|
+
#
|
572
|
+
# @return [Series]
|
573
|
+
def log(base = Math::E)
|
574
|
+
super
|
575
|
+
end
|
576
|
+
|
577
|
+
# Compute the base 10 logarithm of the input array, element-wise.
|
578
|
+
#
|
579
|
+
# @return [Series]
|
580
|
+
def log10
|
581
|
+
super
|
582
|
+
end
|
583
|
+
|
584
|
+
# Compute the exponential, element-wise.
|
585
|
+
#
|
586
|
+
# @return [Series]
|
587
|
+
def exp
|
588
|
+
super
|
589
|
+
end
|
590
|
+
|
591
|
+
# Create a new Series that copies data from this Series without null values.
|
592
|
+
#
|
593
|
+
# @return [Series]
|
594
|
+
def drop_nulls
|
595
|
+
super
|
596
|
+
end
|
597
|
+
|
598
|
+
# Drop NaN values.
|
599
|
+
#
|
600
|
+
# @return [Series]
|
601
|
+
def drop_nans
|
602
|
+
super
|
603
|
+
end
|
604
|
+
|
605
|
+
# Cast this Series to a DataFrame.
|
606
|
+
#
|
607
|
+
# @return [DataFrame]
|
608
|
+
def to_frame
|
609
|
+
Utils.wrap_df(RbDataFrame.new([_s]))
|
610
|
+
end
|
611
|
+
|
612
|
+
# Quick summary statistics of a series.
|
613
|
+
#
|
614
|
+
# Series with mixed datatypes will return summary statistics for the datatype of
|
615
|
+
# the first value.
|
616
|
+
#
|
617
|
+
# @return [DataFrame]
|
618
|
+
#
|
619
|
+
# @example
|
620
|
+
# series_num = Polars::Series.new([1, 2, 3, 4, 5])
|
621
|
+
# series_num.describe
|
622
|
+
# # =>
|
623
|
+
# # shape: (6, 2)
|
624
|
+
# # ┌────────────┬──────────┐
|
625
|
+
# # │ statistic ┆ value │
|
626
|
+
# # │ --- ┆ --- │
|
627
|
+
# # │ str ┆ f64 │
|
628
|
+
# # ╞════════════╪══════════╡
|
629
|
+
# # │ min ┆ 1.0 │
|
630
|
+
# # │ max ┆ 5.0 │
|
631
|
+
# # │ null_count ┆ 0.0 │
|
632
|
+
# # │ mean ┆ 3.0 │
|
633
|
+
# # │ std ┆ 1.581139 │
|
634
|
+
# # │ count ┆ 5.0 │
|
635
|
+
# # └────────────┴──────────┘
|
636
|
+
#
|
637
|
+
# @example
|
638
|
+
# series_str = Polars::Series.new(["a", "a", nil, "b", "c"])
|
639
|
+
# series_str.describe
|
640
|
+
# # =>
|
641
|
+
# # shape: (3, 2)
|
642
|
+
# # ┌────────────┬───────┐
|
643
|
+
# # │ statistic ┆ value │
|
644
|
+
# # │ --- ┆ --- │
|
645
|
+
# # │ str ┆ i64 │
|
646
|
+
# # ╞════════════╪═══════╡
|
647
|
+
# # │ unique ┆ 4 │
|
648
|
+
# # │ null_count ┆ 1 │
|
649
|
+
# # │ count ┆ 5 │
|
650
|
+
# # └────────────┴───────┘
|
651
|
+
def describe
|
652
|
+
if len == 0
|
653
|
+
raise ArgumentError, "Series must contain at least one value"
|
654
|
+
elsif is_numeric
|
655
|
+
s = cast(:f64)
|
656
|
+
stats = {
|
657
|
+
"min" => s.min,
|
658
|
+
"max" => s.max,
|
659
|
+
"null_count" => s.null_count,
|
660
|
+
"mean" => s.mean,
|
661
|
+
"std" => s.std,
|
662
|
+
"count" => s.len
|
663
|
+
}
|
664
|
+
elsif is_boolean
|
665
|
+
stats = {
|
666
|
+
"sum" => sum,
|
667
|
+
"null_count" => null_count,
|
668
|
+
"count" => len
|
669
|
+
}
|
670
|
+
elsif is_utf8
|
671
|
+
stats = {
|
672
|
+
"unique" => unique.length,
|
673
|
+
"null_count" => null_count,
|
674
|
+
"count" => len
|
675
|
+
}
|
676
|
+
elsif is_datelike
|
677
|
+
# we coerce all to string, because a polars column
|
678
|
+
# only has a single dtype and dates: datetime and count: int don't match
|
679
|
+
stats = {
|
680
|
+
"min" => dt.min.to_s,
|
681
|
+
"max" => dt.max.to_s,
|
682
|
+
"null_count" => null_count.to_s,
|
683
|
+
"count" => len.to_s
|
684
|
+
}
|
685
|
+
else
|
686
|
+
raise TypeError, "This type is not supported"
|
687
|
+
end
|
688
|
+
|
689
|
+
Polars::DataFrame.new(
|
690
|
+
{"statistic" => stats.keys, "value" => stats.values}
|
691
|
+
)
|
692
|
+
end
|
693
|
+
|
694
|
+
# Reduce this Series to the sum value.
|
695
|
+
#
|
696
|
+
# @return [Numeric]
|
697
|
+
#
|
698
|
+
# @note
|
699
|
+
# Dtypes `:i8`, `:u8`, `:i16`, and `:u16` are cast to
|
700
|
+
# `:i64` before summing to prevent overflow issues.
|
701
|
+
#
|
702
|
+
# @example
|
703
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
704
|
+
# s.sum
|
705
|
+
# # => 6
|
706
|
+
def sum
|
707
|
+
_s.sum
|
708
|
+
end
|
709
|
+
|
710
|
+
# Reduce this Series to the mean value.
|
711
|
+
#
|
712
|
+
# @return [Float, nil]
|
713
|
+
#
|
714
|
+
# @example
|
715
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
716
|
+
# s.mean
|
717
|
+
# # => 2.0
|
718
|
+
def mean
|
719
|
+
_s.mean
|
720
|
+
end
|
721
|
+
|
722
|
+
# Reduce this Series to the product value.
|
723
|
+
#
|
724
|
+
# @return [Numeric]
|
725
|
+
def product
|
726
|
+
to_frame.select(Polars.col(name).product).to_series[0]
|
727
|
+
end
|
728
|
+
|
729
|
+
# Get the minimal value in this Series.
|
730
|
+
#
|
731
|
+
# @return [Object]
|
732
|
+
#
|
733
|
+
# @example
|
734
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
735
|
+
# s.min
|
736
|
+
# # => 1
|
737
|
+
def min
|
738
|
+
_s.min
|
739
|
+
end
|
740
|
+
|
741
|
+
# Get the maximum value in this Series.
|
742
|
+
#
|
743
|
+
# @return [Object]
|
744
|
+
#
|
745
|
+
# @example
|
746
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
747
|
+
# s.max
|
748
|
+
# # => 3
|
749
|
+
def max
|
750
|
+
_s.max
|
751
|
+
end
|
752
|
+
|
753
|
+
# Get maximum value, but propagate/poison encountered NaN values.
|
754
|
+
#
|
755
|
+
# @return [Object]
|
756
|
+
def nan_max
|
757
|
+
to_frame.select(Polars.col(name).nan_max)[0, 0]
|
758
|
+
end
|
759
|
+
|
760
|
+
# Get minimum value, but propagate/poison encountered NaN values.
|
761
|
+
#
|
762
|
+
# @return [Object]
|
763
|
+
def nan_min
|
764
|
+
to_frame.select(Polars.col(name).nan_min)[0, 0]
|
765
|
+
end
|
766
|
+
|
767
|
+
# Get the standard deviation of this Series.
|
768
|
+
#
|
769
|
+
# @param ddof [Integer]
|
770
|
+
# “Delta Degrees of Freedom”: the divisor used in the calculation is N - ddof,
|
771
|
+
# where N represents the number of elements.
|
772
|
+
#
|
773
|
+
# @return [Float, nil]
|
774
|
+
#
|
775
|
+
# @example
|
776
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
777
|
+
# s.std
|
778
|
+
# # => 1.0
|
779
|
+
def std(ddof: 1)
|
780
|
+
if !is_numeric
|
781
|
+
nil
|
782
|
+
else
|
783
|
+
to_frame.select(Polars.col(name).std(ddof: ddof)).to_series[0]
|
784
|
+
end
|
785
|
+
end
|
786
|
+
|
787
|
+
# Get variance of this Series.
|
788
|
+
#
|
789
|
+
# @param ddof [Integer]
|
790
|
+
# “Delta Degrees of Freedom”: the divisor used in the calculation is N - ddof,
|
791
|
+
# where N represents the number of elements.
|
792
|
+
#
|
793
|
+
# @return [Float, nil]
|
794
|
+
#
|
795
|
+
# @example
|
796
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
797
|
+
# s.var
|
798
|
+
# # => 1.0
|
799
|
+
def var(ddof: 1)
|
800
|
+
if !is_numeric
|
801
|
+
nil
|
802
|
+
else
|
803
|
+
to_frame.select(Polars.col(name).var(ddof: ddof)).to_series[0]
|
804
|
+
end
|
805
|
+
end
|
806
|
+
|
807
|
+
# Get the median of this Series.
|
808
|
+
#
|
809
|
+
# @return [Float, nil]
|
810
|
+
#
|
811
|
+
# @example
|
812
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
813
|
+
# s.median
|
814
|
+
# # => 2.0
|
815
|
+
def median
|
816
|
+
_s.median
|
817
|
+
end
|
818
|
+
|
819
|
+
# Get the quantile value of this Series.
|
820
|
+
#
|
821
|
+
# @param quantile [Float, nil]
|
822
|
+
# Quantile between 0.0 and 1.0.
|
823
|
+
# @param interpolation ["nearest", "higher", "lower", "midpoint", "linear"]
|
824
|
+
# Interpolation method.
|
825
|
+
#
|
826
|
+
# @return [Float, nil]
|
827
|
+
#
|
828
|
+
# @example
|
829
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
830
|
+
# s.quantile(0.5)
|
831
|
+
# # => 2.0
|
832
|
+
def quantile(quantile, interpolation: "nearest")
|
833
|
+
_s.quantile(quantile, interpolation)
|
834
|
+
end
|
835
|
+
|
836
|
+
# Get dummy variables.
|
837
|
+
#
|
838
|
+
# @return [DataFrame]
|
839
|
+
#
|
840
|
+
# @example
|
841
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
842
|
+
# s.to_dummies
|
843
|
+
# # =>
|
844
|
+
# # shape: (3, 3)
|
845
|
+
# # ┌─────┬─────┬─────┐
|
846
|
+
# # │ a_1 ┆ a_2 ┆ a_3 │
|
847
|
+
# # │ --- ┆ --- ┆ --- │
|
848
|
+
# # │ u8 ┆ u8 ┆ u8 │
|
849
|
+
# # ╞═════╪═════╪═════╡
|
850
|
+
# # │ 1 ┆ 0 ┆ 0 │
|
851
|
+
# # │ 0 ┆ 1 ┆ 0 │
|
852
|
+
# # │ 0 ┆ 0 ┆ 1 │
|
853
|
+
# # └─────┴─────┴─────┘
|
854
|
+
def to_dummies(separator: "_", drop_first: false)
|
855
|
+
Utils.wrap_df(_s.to_dummies(separator, drop_first))
|
856
|
+
end
|
857
|
+
|
858
|
+
# Bin continuous values into discrete categories.
|
859
|
+
#
|
860
|
+
# @param breaks [Array]
|
861
|
+
# List of unique cut points.
|
862
|
+
# @param labels [Array]
|
863
|
+
# Names of the categories. The number of labels must be equal to the number
|
864
|
+
# of cut points plus one.
|
865
|
+
# @param left_closed [Boolean]
|
866
|
+
# Set the intervals to be left-closed instead of right-closed.
|
867
|
+
# @param include_breaks [Boolean]
|
868
|
+
# Include a column with the right endpoint of the bin each observation falls
|
869
|
+
# in. This will change the data type of the output from a
|
870
|
+
# `Categorical` to a `Struct`.
|
871
|
+
#
|
872
|
+
# @return [Series]
|
873
|
+
#
|
874
|
+
# @example Divide the column into three categories.
|
875
|
+
# s = Polars::Series.new("foo", [-2, -1, 0, 1, 2])
|
876
|
+
# s.cut([-1, 1], labels: ["a", "b", "c"])
|
877
|
+
# # =>
|
878
|
+
# # shape: (5,)
|
879
|
+
# # Series: 'foo' [cat]
|
880
|
+
# # [
|
881
|
+
# # "a"
|
882
|
+
# # "a"
|
883
|
+
# # "b"
|
884
|
+
# # "b"
|
885
|
+
# # "c"
|
886
|
+
# # ]
|
887
|
+
#
|
888
|
+
# @example Create a DataFrame with the breakpoint and category for each value.
|
889
|
+
# cut = s.cut([-1, 1], include_breaks: true).alias("cut")
|
890
|
+
# s.to_frame.with_columns(cut).unnest("cut")
|
891
|
+
# # =>
|
892
|
+
# # shape: (5, 3)
|
893
|
+
# # ┌─────┬─────────────┬────────────┐
|
894
|
+
# # │ foo ┆ break_point ┆ category │
|
895
|
+
# # │ --- ┆ --- ┆ --- │
|
896
|
+
# # │ i64 ┆ f64 ┆ cat │
|
897
|
+
# # ╞═════╪═════════════╪════════════╡
|
898
|
+
# # │ -2 ┆ -1.0 ┆ (-inf, -1] │
|
899
|
+
# # │ -1 ┆ -1.0 ┆ (-inf, -1] │
|
900
|
+
# # │ 0 ┆ 1.0 ┆ (-1, 1] │
|
901
|
+
# # │ 1 ┆ 1.0 ┆ (-1, 1] │
|
902
|
+
# # │ 2 ┆ inf ┆ (1, inf] │
|
903
|
+
# # └─────┴─────────────┴────────────┘
|
904
|
+
def cut(breaks, labels: nil, left_closed: false, include_breaks: false)
|
905
|
+
result = (
|
906
|
+
to_frame
|
907
|
+
.select(
|
908
|
+
Polars.col(name).cut(
|
909
|
+
breaks,
|
910
|
+
labels: labels,
|
911
|
+
left_closed: left_closed,
|
912
|
+
include_breaks: include_breaks
|
913
|
+
)
|
914
|
+
)
|
915
|
+
.to_series
|
916
|
+
)
|
917
|
+
|
918
|
+
if include_breaks
|
919
|
+
result = result.struct.rename_fields(["break_point", "category"])
|
920
|
+
end
|
921
|
+
|
922
|
+
result
|
923
|
+
end
|
924
|
+
|
925
|
+
# Bin continuous values into discrete categories based on their quantiles.
|
926
|
+
#
|
927
|
+
# @param quantiles [Array]
|
928
|
+
# Either a list of quantile probabilities between 0 and 1 or a positive
|
929
|
+
# integer determining the number of bins with uniform probability.
|
930
|
+
# @param labels [Array]
|
931
|
+
# Names of the categories. The number of labels must be equal to the number
|
932
|
+
# of cut points plus one.
|
933
|
+
# @param left_closed [Boolean]
|
934
|
+
# Set the intervals to be left-closed instead of right-closed.
|
935
|
+
# @param allow_duplicates [Boolean]
|
936
|
+
# If set to `true`, duplicates in the resulting quantiles are dropped,
|
937
|
+
# rather than raising a `DuplicateError`. This can happen even with unique
|
938
|
+
# probabilities, depending on the data.
|
939
|
+
# @param include_breaks [Boolean]
|
940
|
+
# Include a column with the right endpoint of the bin each observation falls
|
941
|
+
# in. This will change the data type of the output from a
|
942
|
+
# `Categorical` to a `Struct`.
|
943
|
+
#
|
944
|
+
# @return [Series]
|
945
|
+
#
|
946
|
+
# @example Divide a column into three categories according to pre-defined quantile probabilities.
|
947
|
+
# s = Polars::Series.new("foo", [-2, -1, 0, 1, 2])
|
948
|
+
# s.qcut([0.25, 0.75], labels: ["a", "b", "c"])
|
949
|
+
# # =>
|
950
|
+
# # shape: (5,)
|
951
|
+
# # Series: 'foo' [cat]
|
952
|
+
# # [
|
953
|
+
# # "a"
|
954
|
+
# # "a"
|
955
|
+
# # "b"
|
956
|
+
# # "b"
|
957
|
+
# # "c"
|
958
|
+
# # ]
|
959
|
+
#
|
960
|
+
# @example Divide a column into two categories using uniform quantile probabilities.
|
961
|
+
# s.qcut(2, labels: ["low", "high"], left_closed: true)
|
962
|
+
# # =>
|
963
|
+
# # shape: (5,)
|
964
|
+
# # Series: 'foo' [cat]
|
965
|
+
# # [
|
966
|
+
# # "low"
|
967
|
+
# # "low"
|
968
|
+
# # "high"
|
969
|
+
# # "high"
|
970
|
+
# # "high"
|
971
|
+
# # ]
|
972
|
+
#
|
973
|
+
# @example Create a DataFrame with the breakpoint and category for each value.
|
974
|
+
# cut = s.qcut([0.25, 0.75], include_breaks: true).alias("cut")
|
975
|
+
# s.to_frame.with_columns(cut).unnest("cut")
|
976
|
+
# # =>
|
977
|
+
# # shape: (5, 3)
|
978
|
+
# # ┌─────┬─────────────┬────────────┐
|
979
|
+
# # │ foo ┆ break_point ┆ category │
|
980
|
+
# # │ --- ┆ --- ┆ --- │
|
981
|
+
# # │ i64 ┆ f64 ┆ cat │
|
982
|
+
# # ╞═════╪═════════════╪════════════╡
|
983
|
+
# # │ -2 ┆ -1.0 ┆ (-inf, -1] │
|
984
|
+
# # │ -1 ┆ -1.0 ┆ (-inf, -1] │
|
985
|
+
# # │ 0 ┆ 1.0 ┆ (-1, 1] │
|
986
|
+
# # │ 1 ┆ 1.0 ┆ (-1, 1] │
|
987
|
+
# # │ 2 ┆ inf ┆ (1, inf] │
|
988
|
+
# # └─────┴─────────────┴────────────┘
|
989
|
+
def qcut(quantiles, labels: nil, left_closed: false, allow_duplicates: false, include_breaks: false)
|
990
|
+
result = (
|
991
|
+
to_frame
|
992
|
+
.select(
|
993
|
+
Polars.col(name).qcut(
|
994
|
+
quantiles,
|
995
|
+
labels: labels,
|
996
|
+
left_closed: left_closed,
|
997
|
+
allow_duplicates: allow_duplicates,
|
998
|
+
include_breaks: include_breaks
|
999
|
+
)
|
1000
|
+
)
|
1001
|
+
.to_series
|
1002
|
+
)
|
1003
|
+
|
1004
|
+
if include_breaks
|
1005
|
+
result = result.struct.rename_fields(["break_point", "category"])
|
1006
|
+
end
|
1007
|
+
|
1008
|
+
result
|
1009
|
+
end
|
1010
|
+
|
1011
|
+
# Get the lengths of runs of identical values.
|
1012
|
+
#
|
1013
|
+
# @return [Series]
|
1014
|
+
#
|
1015
|
+
# @example
|
1016
|
+
# s = Polars::Series.new("s", [1, 1, 2, 1, nil, 1, 3, 3])
|
1017
|
+
# s.rle.struct.unnest
|
1018
|
+
# # =>
|
1019
|
+
# # shape: (6, 2)
|
1020
|
+
# # ┌─────┬───────┐
|
1021
|
+
# # │ len ┆ value │
|
1022
|
+
# # │ --- ┆ --- │
|
1023
|
+
# # │ u32 ┆ i64 │
|
1024
|
+
# # ╞═════╪═══════╡
|
1025
|
+
# # │ 2 ┆ 1 │
|
1026
|
+
# # │ 1 ┆ 2 │
|
1027
|
+
# # │ 1 ┆ 1 │
|
1028
|
+
# # │ 1 ┆ null │
|
1029
|
+
# # │ 1 ┆ 1 │
|
1030
|
+
# # │ 2 ┆ 3 │
|
1031
|
+
# # └─────┴───────┘
|
1032
|
+
def rle
|
1033
|
+
super
|
1034
|
+
end
|
1035
|
+
|
1036
|
+
# Map values to run IDs.
|
1037
|
+
#
|
1038
|
+
# Similar to RLE, but it maps each value to an ID corresponding to the run into
|
1039
|
+
# which it falls. This is especially useful when you want to define groups by
|
1040
|
+
# runs of identical values rather than the values themselves.
|
1041
|
+
#
|
1042
|
+
# @return [Series]
|
1043
|
+
#
|
1044
|
+
# @example
|
1045
|
+
# s = Polars::Series.new("s", [1, 1, 2, 1, nil, 1, 3, 3])
|
1046
|
+
# s.rle_id
|
1047
|
+
# # =>
|
1048
|
+
# # shape: (8,)
|
1049
|
+
# # Series: 's' [u32]
|
1050
|
+
# # [
|
1051
|
+
# # 0
|
1052
|
+
# # 0
|
1053
|
+
# # 1
|
1054
|
+
# # 2
|
1055
|
+
# # 3
|
1056
|
+
# # 4
|
1057
|
+
# # 5
|
1058
|
+
# # 5
|
1059
|
+
# # ]
|
1060
|
+
def rle_id
|
1061
|
+
super
|
1062
|
+
end
|
1063
|
+
|
1064
|
+
# Count the unique values in a Series.
|
1065
|
+
#
|
1066
|
+
# @param sort [Boolean]
|
1067
|
+
# Ensure the output is sorted from most values to least.
|
1068
|
+
#
|
1069
|
+
# @return [DataFrame]
|
1070
|
+
#
|
1071
|
+
# @example
|
1072
|
+
# s = Polars::Series.new("a", [1, 2, 2, 3])
|
1073
|
+
# s.value_counts.sort("a")
|
1074
|
+
# # =>
|
1075
|
+
# # shape: (3, 2)
|
1076
|
+
# # ┌─────┬────────┐
|
1077
|
+
# # │ a ┆ counts │
|
1078
|
+
# # │ --- ┆ --- │
|
1079
|
+
# # │ i64 ┆ u32 │
|
1080
|
+
# # ╞═════╪════════╡
|
1081
|
+
# # │ 1 ┆ 1 │
|
1082
|
+
# # │ 2 ┆ 2 │
|
1083
|
+
# # │ 3 ┆ 1 │
|
1084
|
+
# # └─────┴────────┘
|
1085
|
+
def value_counts(
|
1086
|
+
sort: false,
|
1087
|
+
parallel: false,
|
1088
|
+
name: nil,
|
1089
|
+
normalize: false
|
1090
|
+
)
|
1091
|
+
if name.nil?
|
1092
|
+
if normalize
|
1093
|
+
name = "proportion"
|
1094
|
+
else
|
1095
|
+
name = "count"
|
1096
|
+
end
|
1097
|
+
end
|
1098
|
+
DataFrame._from_rbdf(
|
1099
|
+
self._s.value_counts(
|
1100
|
+
sort, parallel, name, normalize
|
1101
|
+
)
|
1102
|
+
)
|
1103
|
+
end
|
1104
|
+
|
1105
|
+
# Return a count of the unique values in the order of appearance.
|
1106
|
+
#
|
1107
|
+
# @return [Series]
|
1108
|
+
#
|
1109
|
+
# @example
|
1110
|
+
# s = Polars::Series.new("id", ["a", "b", "b", "c", "c", "c"])
|
1111
|
+
# s.unique_counts
|
1112
|
+
# # =>
|
1113
|
+
# # shape: (3,)
|
1114
|
+
# # Series: 'id' [u32]
|
1115
|
+
# # [
|
1116
|
+
# # 1
|
1117
|
+
# # 2
|
1118
|
+
# # 3
|
1119
|
+
# # ]
|
1120
|
+
def unique_counts
|
1121
|
+
super
|
1122
|
+
end
|
1123
|
+
|
1124
|
+
# Computes the entropy.
|
1125
|
+
#
|
1126
|
+
# Uses the formula `-sum(pk * log(pk)` where `pk` are discrete probabilities.
|
1127
|
+
#
|
1128
|
+
# @param base [Float]
|
1129
|
+
# Given base, defaults to `e`
|
1130
|
+
# @param normalize [Boolean]
|
1131
|
+
# Normalize pk if it doesn't sum to 1.
|
1132
|
+
#
|
1133
|
+
# @return [Float, nil]
|
1134
|
+
#
|
1135
|
+
# @example
|
1136
|
+
# a = Polars::Series.new([0.99, 0.005, 0.005])
|
1137
|
+
# a.entropy(normalize: true)
|
1138
|
+
# # => 0.06293300616044681
|
1139
|
+
#
|
1140
|
+
# @example
|
1141
|
+
# b = Polars::Series.new([0.65, 0.10, 0.25])
|
1142
|
+
# b.entropy(normalize: true)
|
1143
|
+
# # => 0.8568409950394724
|
1144
|
+
def entropy(base: Math::E, normalize: false)
|
1145
|
+
Polars.select(Polars.lit(self).entropy(base: base, normalize: normalize)).to_series[0]
|
1146
|
+
end
|
1147
|
+
|
1148
|
+
# Run an expression over a sliding window that increases `1` slot every iteration.
|
1149
|
+
#
|
1150
|
+
# @param expr [Expr]
|
1151
|
+
# Expression to evaluate
|
1152
|
+
# @param min_periods [Integer]
|
1153
|
+
# Number of valid values there should be in the window before the expression
|
1154
|
+
# is evaluated. valid values = `length - null_count`
|
1155
|
+
# @param parallel [Boolean]
|
1156
|
+
# Run in parallel. Don't do this in a group by or another operation that
|
1157
|
+
# already has much parallelization.
|
1158
|
+
#
|
1159
|
+
# @return [Series]
|
1160
|
+
#
|
1161
|
+
# @note
|
1162
|
+
# This functionality is experimental and may change without it being considered a
|
1163
|
+
# breaking change.
|
1164
|
+
#
|
1165
|
+
# @note
|
1166
|
+
# This can be really slow as it can have `O(n^2)` complexity. Don't use this
|
1167
|
+
# for operations that visit all elements.
|
1168
|
+
#
|
1169
|
+
# @example
|
1170
|
+
# s = Polars::Series.new("values", [1, 2, 3, 4, 5])
|
1171
|
+
# s.cumulative_eval(Polars.element.first - Polars.element.last ** 2)
|
1172
|
+
# # =>
|
1173
|
+
# # shape: (5,)
|
1174
|
+
# # Series: 'values' [i64]
|
1175
|
+
# # [
|
1176
|
+
# # 0
|
1177
|
+
# # -3
|
1178
|
+
# # -8
|
1179
|
+
# # -15
|
1180
|
+
# # -24
|
1181
|
+
# # ]
|
1182
|
+
def cumulative_eval(expr, min_periods: 1, parallel: false)
|
1183
|
+
super
|
1184
|
+
end
|
1185
|
+
|
1186
|
+
# Return a copy of the Series with a new alias/name.
|
1187
|
+
#
|
1188
|
+
# @param name [String]
|
1189
|
+
# New name.
|
1190
|
+
#
|
1191
|
+
# @return [Series]
|
1192
|
+
#
|
1193
|
+
# @example
|
1194
|
+
# s = Polars::Series.new("x", [1, 2, 3])
|
1195
|
+
# s.alias("y")
|
1196
|
+
def alias(name)
|
1197
|
+
s = dup
|
1198
|
+
s._s.rename(name)
|
1199
|
+
s
|
1200
|
+
end
|
1201
|
+
|
1202
|
+
# Rename this Series.
|
1203
|
+
#
|
1204
|
+
# @param name [String]
|
1205
|
+
# New name.
|
1206
|
+
# @param in_place [Boolean]
|
1207
|
+
# Modify the Series in-place.
|
1208
|
+
#
|
1209
|
+
# @return [Series]
|
1210
|
+
#
|
1211
|
+
# @example
|
1212
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1213
|
+
# s.rename("b")
|
1214
|
+
def rename(name, in_place: false)
|
1215
|
+
if in_place
|
1216
|
+
_s.rename(name)
|
1217
|
+
self
|
1218
|
+
else
|
1219
|
+
self.alias(name)
|
1220
|
+
end
|
1221
|
+
end
|
1222
|
+
|
1223
|
+
# Get the length of each individual chunk.
|
1224
|
+
#
|
1225
|
+
# @return [Array]
|
1226
|
+
#
|
1227
|
+
# @example
|
1228
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1229
|
+
# s2 = Polars::Series.new("b", [4, 5, 6])
|
1230
|
+
#
|
1231
|
+
# @example Concatenate Series with rechunk: true
|
1232
|
+
# Polars.concat([s, s2]).chunk_lengths
|
1233
|
+
# # => [6]
|
1234
|
+
#
|
1235
|
+
# @example Concatenate Series with rechunk: false
|
1236
|
+
# Polars.concat([s, s2], rechunk: false).chunk_lengths
|
1237
|
+
# # => [3, 3]
|
1238
|
+
def chunk_lengths
|
1239
|
+
_s.chunk_lengths
|
1240
|
+
end
|
1241
|
+
|
1242
|
+
# Get the number of chunks that this Series contains.
|
1243
|
+
#
|
1244
|
+
# @return [Integer]
|
1245
|
+
#
|
1246
|
+
# @example
|
1247
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1248
|
+
# s2 = Polars::Series.new("b", [4, 5, 6])
|
1249
|
+
#
|
1250
|
+
# @example Concatenate Series with rechunk: true
|
1251
|
+
# Polars.concat([s, s2]).n_chunks
|
1252
|
+
# # => 1
|
1253
|
+
#
|
1254
|
+
# @example Concatenate Series with rechunk: false
|
1255
|
+
# Polars.concat([s, s2], rechunk: false).n_chunks
|
1256
|
+
# # => 2
|
1257
|
+
def n_chunks
|
1258
|
+
_s.n_chunks
|
1259
|
+
end
|
1260
|
+
|
1261
|
+
# Get an array with the cumulative sum computed at every element.
|
1262
|
+
#
|
1263
|
+
# @param reverse [Boolean]
|
1264
|
+
# reverse the operation.
|
1265
|
+
#
|
1266
|
+
# @return [Series]
|
1267
|
+
#
|
1268
|
+
# @note
|
1269
|
+
# Dtypes `:i8`, `:u8`, `:i16`, and `:u16` are cast to
|
1270
|
+
# `:i64` before summing to prevent overflow issues.
|
1271
|
+
#
|
1272
|
+
# @example
|
1273
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1274
|
+
# s.cum_sum
|
1275
|
+
# # =>
|
1276
|
+
# # shape: (3,)
|
1277
|
+
# # Series: 'a' [i64]
|
1278
|
+
# # [
|
1279
|
+
# # 1
|
1280
|
+
# # 3
|
1281
|
+
# # 6
|
1282
|
+
# # ]
|
1283
|
+
def cum_sum(reverse: false)
|
1284
|
+
super
|
1285
|
+
end
|
1286
|
+
alias_method :cumsum, :cum_sum
|
1287
|
+
|
1288
|
+
# Get an array with the cumulative min computed at every element.
|
1289
|
+
#
|
1290
|
+
# @param reverse [Boolean]
|
1291
|
+
# reverse the operation.
|
1292
|
+
#
|
1293
|
+
# @return [Series]
|
1294
|
+
#
|
1295
|
+
# @example
|
1296
|
+
# s = Polars::Series.new("a", [3, 5, 1])
|
1297
|
+
# s.cum_min
|
1298
|
+
# # =>
|
1299
|
+
# # shape: (3,)
|
1300
|
+
# # Series: 'a' [i64]
|
1301
|
+
# # [
|
1302
|
+
# # 3
|
1303
|
+
# # 3
|
1304
|
+
# # 1
|
1305
|
+
# # ]
|
1306
|
+
def cum_min(reverse: false)
|
1307
|
+
super
|
1308
|
+
end
|
1309
|
+
alias_method :cummin, :cum_min
|
1310
|
+
|
1311
|
+
# Get an array with the cumulative max computed at every element.
|
1312
|
+
#
|
1313
|
+
# @param reverse [Boolean]
|
1314
|
+
# reverse the operation.
|
1315
|
+
#
|
1316
|
+
# @return [Series]
|
1317
|
+
#
|
1318
|
+
# @example
|
1319
|
+
# s = Polars::Series.new("a", [3, 5, 1])
|
1320
|
+
# s.cum_max
|
1321
|
+
# # =>
|
1322
|
+
# # shape: (3,)
|
1323
|
+
# # Series: 'a' [i64]
|
1324
|
+
# # [
|
1325
|
+
# # 3
|
1326
|
+
# # 5
|
1327
|
+
# # 5
|
1328
|
+
# # ]
|
1329
|
+
def cum_max(reverse: false)
|
1330
|
+
super
|
1331
|
+
end
|
1332
|
+
alias_method :cummax, :cum_max
|
1333
|
+
|
1334
|
+
# Get an array with the cumulative product computed at every element.
|
1335
|
+
#
|
1336
|
+
# @param reverse [Boolean]
|
1337
|
+
# reverse the operation.
|
1338
|
+
#
|
1339
|
+
# @return [Series]
|
1340
|
+
#
|
1341
|
+
# @note
|
1342
|
+
# Dtypes `:i8`, `:u8`, `:i16`, and `:u16` are cast to
|
1343
|
+
# `:i64` before multiplying to prevent overflow issues.
|
1344
|
+
#
|
1345
|
+
# @example
|
1346
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1347
|
+
# s.cum_prod
|
1348
|
+
# # =>
|
1349
|
+
# # shape: (3,)
|
1350
|
+
# # Series: 'a' [i64]
|
1351
|
+
# # [
|
1352
|
+
# # 1
|
1353
|
+
# # 2
|
1354
|
+
# # 6
|
1355
|
+
# # ]
|
1356
|
+
def cum_prod(reverse: false)
|
1357
|
+
super
|
1358
|
+
end
|
1359
|
+
alias_method :cumprod, :cum_prod
|
1360
|
+
|
1361
|
+
# Get the first `n` rows.
|
1362
|
+
#
|
1363
|
+
# Alias for {#head}.
|
1364
|
+
#
|
1365
|
+
# @param n [Integer]
|
1366
|
+
# Number of rows to return.
|
1367
|
+
#
|
1368
|
+
# @return [Series]
|
1369
|
+
#
|
1370
|
+
# @example
|
1371
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1372
|
+
# s.limit(2)
|
1373
|
+
# # =>
|
1374
|
+
# # shape: (2,)
|
1375
|
+
# # Series: 'a' [i64]
|
1376
|
+
# # [
|
1377
|
+
# # 1
|
1378
|
+
# # 2
|
1379
|
+
# # ]
|
1380
|
+
def limit(n = 10)
|
1381
|
+
to_frame.select(F.col(name).limit(n)).to_series
|
1382
|
+
end
|
1383
|
+
|
1384
|
+
# Get a slice of this Series.
|
1385
|
+
#
|
1386
|
+
# @param offset [Integer]
|
1387
|
+
# Start index. Negative indexing is supported.
|
1388
|
+
# @param length [Integer, nil]
|
1389
|
+
# Length of the slice. If set to `nil`, all rows starting at the offset
|
1390
|
+
# will be selected.
|
1391
|
+
#
|
1392
|
+
# @return [Series]
|
1393
|
+
#
|
1394
|
+
# @example
|
1395
|
+
# s = Polars::Series.new("a", [1, 2, 3, 4])
|
1396
|
+
# s.slice(1, 2)
|
1397
|
+
# # =>
|
1398
|
+
# # shape: (2,)
|
1399
|
+
# # Series: 'a' [i64]
|
1400
|
+
# # [
|
1401
|
+
# # 2
|
1402
|
+
# # 3
|
1403
|
+
# # ]
|
1404
|
+
def slice(offset, length = nil)
|
1405
|
+
self.class._from_rbseries(_s.slice(offset, length))
|
1406
|
+
end
|
1407
|
+
|
1408
|
+
# Append a Series to this one.
|
1409
|
+
#
|
1410
|
+
# @param other [Series]
|
1411
|
+
# Series to append.
|
1412
|
+
# @param append_chunks [Boolean]
|
1413
|
+
# If set to `true` the append operation will add the chunks from `other` to
|
1414
|
+
# self. This is super cheap.
|
1415
|
+
#
|
1416
|
+
# If set to `false` the append operation will do the same as
|
1417
|
+
# {DataFrame#extend} which extends the memory backed by this Series with
|
1418
|
+
# the values from `other`.
|
1419
|
+
#
|
1420
|
+
# Different from `append_chunks`, `extend` appends the data from `other` to
|
1421
|
+
# the underlying memory locations and thus may cause a reallocation (which is
|
1422
|
+
# expensive).
|
1423
|
+
#
|
1424
|
+
# If this does not cause a reallocation, the resulting data structure will not
|
1425
|
+
# have any extra chunks and thus will yield faster queries.
|
1426
|
+
#
|
1427
|
+
# Prefer `extend` over `append_chunks` when you want to do a query after a
|
1428
|
+
# single append. For instance during online operations where you add `n` rows
|
1429
|
+
# and rerun a query.
|
1430
|
+
#
|
1431
|
+
# Prefer `append_chunks` over `extend` when you want to append many times
|
1432
|
+
# before doing a query. For instance, when you read in multiple files and when
|
1433
|
+
# to store them in a single Series. In the latter case, finish the sequence
|
1434
|
+
# of `append_chunks` operations with a `rechunk`.
|
1435
|
+
#
|
1436
|
+
# @return [Series]
|
1437
|
+
#
|
1438
|
+
# @example
|
1439
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1440
|
+
# s2 = Polars::Series.new("b", [4, 5, 6])
|
1441
|
+
# s.append(s2)
|
1442
|
+
# # =>
|
1443
|
+
# # shape: (6,)
|
1444
|
+
# # Series: 'a' [i64]
|
1445
|
+
# # [
|
1446
|
+
# # 1
|
1447
|
+
# # 2
|
1448
|
+
# # 3
|
1449
|
+
# # 4
|
1450
|
+
# # 5
|
1451
|
+
# # 6
|
1452
|
+
# # ]
|
1453
|
+
def append(other, append_chunks: true)
|
1454
|
+
begin
|
1455
|
+
if append_chunks
|
1456
|
+
_s.append(other._s)
|
1457
|
+
else
|
1458
|
+
_s.extend(other._s)
|
1459
|
+
end
|
1460
|
+
rescue => e
|
1461
|
+
if e.message == "Already mutably borrowed"
|
1462
|
+
append(other.clone, append_chunks)
|
1463
|
+
else
|
1464
|
+
raise e
|
1465
|
+
end
|
1466
|
+
end
|
1467
|
+
self
|
1468
|
+
end
|
1469
|
+
|
1470
|
+
# Filter elements by a boolean mask.
|
1471
|
+
#
|
1472
|
+
# @param predicate [Series, Array]
|
1473
|
+
# Boolean mask.
|
1474
|
+
#
|
1475
|
+
# @return [Series]
|
1476
|
+
#
|
1477
|
+
# @example
|
1478
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1479
|
+
# mask = Polars::Series.new("", [true, false, true])
|
1480
|
+
# s.filter(mask)
|
1481
|
+
# # =>
|
1482
|
+
# # shape: (2,)
|
1483
|
+
# # Series: 'a' [i64]
|
1484
|
+
# # [
|
1485
|
+
# # 1
|
1486
|
+
# # 3
|
1487
|
+
# # ]
|
1488
|
+
def filter(predicate)
|
1489
|
+
if predicate.is_a?(::Array)
|
1490
|
+
predicate = Series.new("", predicate)
|
1491
|
+
end
|
1492
|
+
Utils.wrap_s(_s.filter(predicate._s))
|
1493
|
+
end
|
1494
|
+
|
1495
|
+
# Get the first `n` rows.
|
1496
|
+
#
|
1497
|
+
# @param n [Integer]
|
1498
|
+
# Number of rows to return.
|
1499
|
+
#
|
1500
|
+
# @return [Series]
|
1501
|
+
#
|
1502
|
+
# @example
|
1503
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1504
|
+
# s.head(2)
|
1505
|
+
# # =>
|
1506
|
+
# # shape: (2,)
|
1507
|
+
# # Series: 'a' [i64]
|
1508
|
+
# # [
|
1509
|
+
# # 1
|
1510
|
+
# # 2
|
1511
|
+
# # ]
|
1512
|
+
def head(n = 10)
|
1513
|
+
to_frame.select(F.col(name).head(n)).to_series
|
1514
|
+
end
|
1515
|
+
|
1516
|
+
# Get the last `n` rows.
|
1517
|
+
#
|
1518
|
+
# @param n [Integer]
|
1519
|
+
# Number of rows to return.
|
1520
|
+
#
|
1521
|
+
# @return [Series]
|
1522
|
+
#
|
1523
|
+
# @example
|
1524
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1525
|
+
# s.tail(2)
|
1526
|
+
# # =>
|
1527
|
+
# # shape: (2,)
|
1528
|
+
# # Series: 'a' [i64]
|
1529
|
+
# # [
|
1530
|
+
# # 2
|
1531
|
+
# # 3
|
1532
|
+
# # ]
|
1533
|
+
def tail(n = 10)
|
1534
|
+
to_frame.select(F.col(name).tail(n)).to_series
|
1535
|
+
end
|
1536
|
+
|
1537
|
+
# Take every nth value in the Series and return as new Series.
|
1538
|
+
#
|
1539
|
+
# @return [Series]
|
1540
|
+
#
|
1541
|
+
# @example
|
1542
|
+
# s = Polars::Series.new("a", [1, 2, 3, 4])
|
1543
|
+
# s.take_every(2)
|
1544
|
+
# # =>
|
1545
|
+
# # shape: (2,)
|
1546
|
+
# # Series: 'a' [i64]
|
1547
|
+
# # [
|
1548
|
+
# # 1
|
1549
|
+
# # 3
|
1550
|
+
# # ]
|
1551
|
+
def take_every(n)
|
1552
|
+
super
|
1553
|
+
end
|
1554
|
+
|
1555
|
+
# Sort this Series.
|
1556
|
+
#
|
1557
|
+
# @param reverse [Boolean]
|
1558
|
+
# Reverse sort.
|
1559
|
+
# @param in_place [Boolean]
|
1560
|
+
# Sort in place.
|
1561
|
+
#
|
1562
|
+
# @return [Series]
|
1563
|
+
#
|
1564
|
+
# @example
|
1565
|
+
# s = Polars::Series.new("a", [1, 3, 4, 2])
|
1566
|
+
# s.sort
|
1567
|
+
# # =>
|
1568
|
+
# # shape: (4,)
|
1569
|
+
# # Series: 'a' [i64]
|
1570
|
+
# # [
|
1571
|
+
# # 1
|
1572
|
+
# # 2
|
1573
|
+
# # 3
|
1574
|
+
# # 4
|
1575
|
+
# # ]
|
1576
|
+
# s.sort(reverse: true)
|
1577
|
+
# # =>
|
1578
|
+
# # shape: (4,)
|
1579
|
+
# # Series: 'a' [i64]
|
1580
|
+
# # [
|
1581
|
+
# # 4
|
1582
|
+
# # 3
|
1583
|
+
# # 2
|
1584
|
+
# # 1
|
1585
|
+
# # ]
|
1586
|
+
def sort(reverse: false, nulls_last: false, multithreaded: true, in_place: false)
|
1587
|
+
if in_place
|
1588
|
+
self._s = _s.sort(reverse, nulls_last, multithreaded)
|
1589
|
+
self
|
1590
|
+
else
|
1591
|
+
Utils.wrap_s(_s.sort(reverse, nulls_last, multithreaded))
|
1592
|
+
end
|
1593
|
+
end
|
1594
|
+
|
1595
|
+
# Return the `k` largest elements.
|
1596
|
+
#
|
1597
|
+
# @param k [Integer]
|
1598
|
+
# Number of elements to return.
|
1599
|
+
#
|
1600
|
+
# @return [Boolean]
|
1601
|
+
#
|
1602
|
+
# @example
|
1603
|
+
# s = Polars::Series.new("a", [2, 5, 1, 4, 3])
|
1604
|
+
# s.top_k(k: 3)
|
1605
|
+
# # =>
|
1606
|
+
# # shape: (3,)
|
1607
|
+
# # Series: 'a' [i64]
|
1608
|
+
# # [
|
1609
|
+
# # 5
|
1610
|
+
# # 4
|
1611
|
+
# # 3
|
1612
|
+
# # ]
|
1613
|
+
def top_k(k: 5)
|
1614
|
+
super
|
1615
|
+
end
|
1616
|
+
|
1617
|
+
# Return the `k` smallest elements.
|
1618
|
+
#
|
1619
|
+
# @param k [Integer]
|
1620
|
+
# Number of elements to return.
|
1621
|
+
#
|
1622
|
+
# @return [Boolean]
|
1623
|
+
#
|
1624
|
+
# @example
|
1625
|
+
# s = Polars::Series.new("a", [2, 5, 1, 4, 3])
|
1626
|
+
# s.bottom_k(k: 3)
|
1627
|
+
# # =>
|
1628
|
+
# # shape: (3,)
|
1629
|
+
# # Series: 'a' [i64]
|
1630
|
+
# # [
|
1631
|
+
# # 1
|
1632
|
+
# # 2
|
1633
|
+
# # 3
|
1634
|
+
# # ]
|
1635
|
+
def bottom_k(k: 5)
|
1636
|
+
super
|
1637
|
+
end
|
1638
|
+
|
1639
|
+
# Get the index values that would sort this Series.
|
1640
|
+
#
|
1641
|
+
# @param reverse [Boolean]
|
1642
|
+
# Sort in reverse (descending) order.
|
1643
|
+
# @param nulls_last [Boolean]
|
1644
|
+
# Place null values last instead of first.
|
1645
|
+
#
|
1646
|
+
# @return [Series]
|
1647
|
+
#
|
1648
|
+
# @example
|
1649
|
+
# s = Polars::Series.new("a", [5, 3, 4, 1, 2])
|
1650
|
+
# s.arg_sort
|
1651
|
+
# # =>
|
1652
|
+
# # shape: (5,)
|
1653
|
+
# # Series: 'a' [u32]
|
1654
|
+
# # [
|
1655
|
+
# # 3
|
1656
|
+
# # 4
|
1657
|
+
# # 1
|
1658
|
+
# # 2
|
1659
|
+
# # 0
|
1660
|
+
# # ]
|
1661
|
+
def arg_sort(reverse: false, nulls_last: false)
|
1662
|
+
super
|
1663
|
+
end
|
1664
|
+
|
1665
|
+
# Get the index values that would sort this Series.
|
1666
|
+
#
|
1667
|
+
# Alias for {#arg_sort}.
|
1668
|
+
#
|
1669
|
+
# @param reverse [Boolean]
|
1670
|
+
# Sort in reverse (descending) order.
|
1671
|
+
# @param nulls_last [Boolean]
|
1672
|
+
# Place null values last instead of first.
|
1673
|
+
#
|
1674
|
+
# @return [Series]
|
1675
|
+
def argsort(reverse: false, nulls_last: false)
|
1676
|
+
super
|
1677
|
+
end
|
1678
|
+
|
1679
|
+
# Get unique index as Series.
|
1680
|
+
#
|
1681
|
+
# @return [Series]
|
1682
|
+
#
|
1683
|
+
# @example
|
1684
|
+
# s = Polars::Series.new("a", [1, 2, 2, 3])
|
1685
|
+
# s.arg_unique
|
1686
|
+
# # =>
|
1687
|
+
# # shape: (3,)
|
1688
|
+
# # Series: 'a' [u32]
|
1689
|
+
# # [
|
1690
|
+
# # 0
|
1691
|
+
# # 1
|
1692
|
+
# # 3
|
1693
|
+
# # ]
|
1694
|
+
def arg_unique
|
1695
|
+
super
|
1696
|
+
end
|
1697
|
+
|
1698
|
+
# Get the index of the minimal value.
|
1699
|
+
#
|
1700
|
+
# @return [Integer, nil]
|
1701
|
+
#
|
1702
|
+
# @example
|
1703
|
+
# s = Polars::Series.new("a", [3, 2, 1])
|
1704
|
+
# s.arg_min
|
1705
|
+
# # => 2
|
1706
|
+
def arg_min
|
1707
|
+
_s.arg_min
|
1708
|
+
end
|
1709
|
+
|
1710
|
+
# Get the index of the maximal value.
|
1711
|
+
#
|
1712
|
+
# @return [Integer, nil]
|
1713
|
+
#
|
1714
|
+
# @example
|
1715
|
+
# s = Polars::Series.new("a", [3, 2, 1])
|
1716
|
+
# s.arg_max
|
1717
|
+
# # => 0
|
1718
|
+
def arg_max
|
1719
|
+
_s.arg_max
|
1720
|
+
end
|
1721
|
+
|
1722
|
+
# Find indices where elements should be inserted to maintain order.
|
1723
|
+
#
|
1724
|
+
# @param element [Object]
|
1725
|
+
# Expression or scalar value.
|
1726
|
+
#
|
1727
|
+
# @return [Integer]
|
1728
|
+
def search_sorted(element, side: "any")
|
1729
|
+
if element.is_a?(Integer) || element.is_a?(Float)
|
1730
|
+
return Polars.select(Polars.lit(self).search_sorted(element, side: side)).item
|
1731
|
+
end
|
1732
|
+
element = Series.new(element)
|
1733
|
+
Polars.select(Polars.lit(self).search_sorted(element, side: side)).to_series
|
1734
|
+
end
|
1735
|
+
|
1736
|
+
# Get unique elements in series.
|
1737
|
+
#
|
1738
|
+
# @param maintain_order [Boolean]
|
1739
|
+
# Maintain order of data. This requires more work.
|
1740
|
+
#
|
1741
|
+
# @return [Series]
|
1742
|
+
#
|
1743
|
+
# @example
|
1744
|
+
# s = Polars::Series.new("a", [1, 2, 2, 3])
|
1745
|
+
# s.unique.sort
|
1746
|
+
# # =>
|
1747
|
+
# # shape: (3,)
|
1748
|
+
# # Series: 'a' [i64]
|
1749
|
+
# # [
|
1750
|
+
# # 1
|
1751
|
+
# # 2
|
1752
|
+
# # 3
|
1753
|
+
# # ]
|
1754
|
+
def unique(maintain_order: false)
|
1755
|
+
super
|
1756
|
+
end
|
1757
|
+
alias_method :uniq, :unique
|
1758
|
+
|
1759
|
+
# Take values by index.
|
1760
|
+
#
|
1761
|
+
# @param indices [Array]
|
1762
|
+
# Index location used for selection.
|
1763
|
+
#
|
1764
|
+
# @return [Series]
|
1765
|
+
#
|
1766
|
+
# @example
|
1767
|
+
# s = Polars::Series.new("a", [1, 2, 3, 4])
|
1768
|
+
# s.take([1, 3])
|
1769
|
+
# # =>
|
1770
|
+
# # shape: (2,)
|
1771
|
+
# # Series: 'a' [i64]
|
1772
|
+
# # [
|
1773
|
+
# # 2
|
1774
|
+
# # 4
|
1775
|
+
# # ]
|
1776
|
+
def take(indices)
|
1777
|
+
to_frame.select(Polars.col(name).take(indices)).to_series
|
1778
|
+
end
|
1779
|
+
|
1780
|
+
# Count the null values in this Series.
|
1781
|
+
#
|
1782
|
+
# @return [Integer]
|
1783
|
+
def null_count
|
1784
|
+
_s.null_count
|
1785
|
+
end
|
1786
|
+
|
1787
|
+
# Return `true` if the Series has a validity bitmask.
|
1788
|
+
#
|
1789
|
+
# If there is none, it means that there are no null values.
|
1790
|
+
# Use this to swiftly assert a Series does not have null values.
|
1791
|
+
#
|
1792
|
+
# @return [Boolean]
|
1793
|
+
def has_nulls
|
1794
|
+
_s.has_nulls
|
1795
|
+
end
|
1796
|
+
alias_method :has_validity, :has_nulls
|
1797
|
+
|
1798
|
+
# Check if the Series is empty.
|
1799
|
+
#
|
1800
|
+
# @return [Boolean]
|
1801
|
+
#
|
1802
|
+
# @example
|
1803
|
+
# s = Polars::Series.new("a", [])
|
1804
|
+
# s.is_empty
|
1805
|
+
# # => true
|
1806
|
+
def is_empty
|
1807
|
+
len == 0
|
1808
|
+
end
|
1809
|
+
alias_method :empty?, :is_empty
|
1810
|
+
|
1811
|
+
# Returns a boolean Series indicating which values are null.
|
1812
|
+
#
|
1813
|
+
# @return [Series]
|
1814
|
+
#
|
1815
|
+
# @example
|
1816
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, nil])
|
1817
|
+
# s.is_null
|
1818
|
+
# # =>
|
1819
|
+
# # shape: (4,)
|
1820
|
+
# # Series: 'a' [bool]
|
1821
|
+
# # [
|
1822
|
+
# # false
|
1823
|
+
# # false
|
1824
|
+
# # false
|
1825
|
+
# # true
|
1826
|
+
# # ]
|
1827
|
+
def is_null
|
1828
|
+
super
|
1829
|
+
end
|
1830
|
+
|
1831
|
+
# Returns a boolean Series indicating which values are not null.
|
1832
|
+
#
|
1833
|
+
# @return [Series]
|
1834
|
+
#
|
1835
|
+
# @example
|
1836
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, nil])
|
1837
|
+
# s.is_not_null
|
1838
|
+
# # =>
|
1839
|
+
# # shape: (4,)
|
1840
|
+
# # Series: 'a' [bool]
|
1841
|
+
# # [
|
1842
|
+
# # true
|
1843
|
+
# # true
|
1844
|
+
# # true
|
1845
|
+
# # false
|
1846
|
+
# # ]
|
1847
|
+
def is_not_null
|
1848
|
+
super
|
1849
|
+
end
|
1850
|
+
|
1851
|
+
# Returns a boolean Series indicating which values are finite.
|
1852
|
+
#
|
1853
|
+
# @return [Series]
|
1854
|
+
#
|
1855
|
+
# @example
|
1856
|
+
# s = Polars::Series.new("a", [1.0, 2.0, Float::INFINITY])
|
1857
|
+
# s.is_finite
|
1858
|
+
# # =>
|
1859
|
+
# # shape: (3,)
|
1860
|
+
# # Series: 'a' [bool]
|
1861
|
+
# # [
|
1862
|
+
# # true
|
1863
|
+
# # true
|
1864
|
+
# # false
|
1865
|
+
# # ]
|
1866
|
+
def is_finite
|
1867
|
+
super
|
1868
|
+
end
|
1869
|
+
|
1870
|
+
# Returns a boolean Series indicating which values are infinite.
|
1871
|
+
#
|
1872
|
+
# @return [Series]
|
1873
|
+
#
|
1874
|
+
# @example
|
1875
|
+
# s = Polars::Series.new("a", [1.0, 2.0, Float::INFINITY])
|
1876
|
+
# s.is_infinite
|
1877
|
+
# # =>
|
1878
|
+
# # shape: (3,)
|
1879
|
+
# # Series: 'a' [bool]
|
1880
|
+
# # [
|
1881
|
+
# # false
|
1882
|
+
# # false
|
1883
|
+
# # true
|
1884
|
+
# # ]
|
1885
|
+
def is_infinite
|
1886
|
+
super
|
1887
|
+
end
|
1888
|
+
|
1889
|
+
# Returns a boolean Series indicating which values are NaN.
|
1890
|
+
#
|
1891
|
+
# @return [Series]
|
1892
|
+
#
|
1893
|
+
# @example
|
1894
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, Float::NAN])
|
1895
|
+
# s.is_nan
|
1896
|
+
# # =>
|
1897
|
+
# # shape: (4,)
|
1898
|
+
# # Series: 'a' [bool]
|
1899
|
+
# # [
|
1900
|
+
# # false
|
1901
|
+
# # false
|
1902
|
+
# # false
|
1903
|
+
# # true
|
1904
|
+
# # ]
|
1905
|
+
def is_nan
|
1906
|
+
super
|
1907
|
+
end
|
1908
|
+
|
1909
|
+
# Returns a boolean Series indicating which values are not NaN.
|
1910
|
+
#
|
1911
|
+
# @return [Series]
|
1912
|
+
#
|
1913
|
+
# @example
|
1914
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, Float::NAN])
|
1915
|
+
# s.is_not_nan
|
1916
|
+
# # =>
|
1917
|
+
# # shape: (4,)
|
1918
|
+
# # Series: 'a' [bool]
|
1919
|
+
# # [
|
1920
|
+
# # true
|
1921
|
+
# # true
|
1922
|
+
# # true
|
1923
|
+
# # false
|
1924
|
+
# # ]
|
1925
|
+
def is_not_nan
|
1926
|
+
super
|
1927
|
+
end
|
1928
|
+
|
1929
|
+
# Check if elements of this Series are in the other Series.
|
1930
|
+
#
|
1931
|
+
# @return [Series]
|
1932
|
+
#
|
1933
|
+
# @example
|
1934
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1935
|
+
# s2 = Polars::Series.new("b", [2, 4])
|
1936
|
+
# s2.is_in(s)
|
1937
|
+
# # =>
|
1938
|
+
# # shape: (2,)
|
1939
|
+
# # Series: 'b' [bool]
|
1940
|
+
# # [
|
1941
|
+
# # true
|
1942
|
+
# # false
|
1943
|
+
# # ]
|
1944
|
+
#
|
1945
|
+
# @example
|
1946
|
+
# sets = Polars::Series.new("sets", [[1, 2, 3], [1, 2], [9, 10]])
|
1947
|
+
# # =>
|
1948
|
+
# # shape: (3,)
|
1949
|
+
# # Series: 'sets' [list[i64]]
|
1950
|
+
# # [
|
1951
|
+
# # [1, 2, 3]
|
1952
|
+
# # [1, 2]
|
1953
|
+
# # [9, 10]
|
1954
|
+
# # ]
|
1955
|
+
#
|
1956
|
+
# @example
|
1957
|
+
# optional_members = Polars::Series.new("optional_members", [1, 2, 3])
|
1958
|
+
# # =>
|
1959
|
+
# # shape: (3,)
|
1960
|
+
# # Series: 'optional_members' [i64]
|
1961
|
+
# # [
|
1962
|
+
# # 1
|
1963
|
+
# # 2
|
1964
|
+
# # 3
|
1965
|
+
# # ]
|
1966
|
+
#
|
1967
|
+
# @example
|
1968
|
+
# optional_members.is_in(sets)
|
1969
|
+
# # =>
|
1970
|
+
# # shape: (3,)
|
1971
|
+
# # Series: 'optional_members' [bool]
|
1972
|
+
# # [
|
1973
|
+
# # true
|
1974
|
+
# # true
|
1975
|
+
# # false
|
1976
|
+
# # ]
|
1977
|
+
def is_in(other)
|
1978
|
+
super
|
1979
|
+
end
|
1980
|
+
alias_method :in?, :is_in
|
1981
|
+
|
1982
|
+
# Get index values where Boolean Series evaluate `true`.
|
1983
|
+
#
|
1984
|
+
# @return [Series]
|
1985
|
+
#
|
1986
|
+
# @example
|
1987
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
1988
|
+
# (s == 2).arg_true
|
1989
|
+
# # =>
|
1990
|
+
# # shape: (1,)
|
1991
|
+
# # Series: 'a' [u32]
|
1992
|
+
# # [
|
1993
|
+
# # 1
|
1994
|
+
# # ]
|
1995
|
+
def arg_true
|
1996
|
+
Polars.arg_where(self, eager: true)
|
1997
|
+
end
|
1998
|
+
|
1999
|
+
# Get mask of all unique values.
|
2000
|
+
#
|
2001
|
+
# @return [Series]
|
2002
|
+
#
|
2003
|
+
# @example
|
2004
|
+
# s = Polars::Series.new("a", [1, 2, 2, 3])
|
2005
|
+
# s.is_unique
|
2006
|
+
# # =>
|
2007
|
+
# # shape: (4,)
|
2008
|
+
# # Series: 'a' [bool]
|
2009
|
+
# # [
|
2010
|
+
# # true
|
2011
|
+
# # false
|
2012
|
+
# # false
|
2013
|
+
# # true
|
2014
|
+
# # ]
|
2015
|
+
def is_unique
|
2016
|
+
super
|
2017
|
+
end
|
2018
|
+
|
2019
|
+
# Get a mask of the first unique value.
|
2020
|
+
#
|
2021
|
+
# @return [Series]
|
2022
|
+
def is_first
|
2023
|
+
super
|
2024
|
+
end
|
2025
|
+
|
2026
|
+
# Get mask of all duplicated values.
|
2027
|
+
#
|
2028
|
+
# @return [Series]
|
2029
|
+
#
|
2030
|
+
# @example
|
2031
|
+
# s = Polars::Series.new("a", [1, 2, 2, 3])
|
2032
|
+
# s.is_duplicated
|
2033
|
+
# # =>
|
2034
|
+
# # shape: (4,)
|
2035
|
+
# # Series: 'a' [bool]
|
2036
|
+
# # [
|
2037
|
+
# # false
|
2038
|
+
# # true
|
2039
|
+
# # true
|
2040
|
+
# # false
|
2041
|
+
# # ]
|
2042
|
+
def is_duplicated
|
2043
|
+
super
|
2044
|
+
end
|
2045
|
+
|
2046
|
+
# Explode a list or utf8 Series.
|
2047
|
+
#
|
2048
|
+
# This means that every item is expanded to a new row.
|
2049
|
+
#
|
2050
|
+
# @return [Series]
|
2051
|
+
#
|
2052
|
+
# @example
|
2053
|
+
# s = Polars::Series.new("a", [[1, 2], [3, 4], [9, 10]])
|
2054
|
+
# s.explode
|
2055
|
+
# # =>
|
2056
|
+
# # shape: (6,)
|
2057
|
+
# # Series: 'a' [i64]
|
2058
|
+
# # [
|
2059
|
+
# # 1
|
2060
|
+
# # 2
|
2061
|
+
# # 3
|
2062
|
+
# # 4
|
2063
|
+
# # 9
|
2064
|
+
# # 10
|
2065
|
+
# # ]
|
2066
|
+
def explode
|
2067
|
+
super
|
2068
|
+
end
|
2069
|
+
|
2070
|
+
# Check if series is equal with another Series.
|
2071
|
+
#
|
2072
|
+
# @param other [Series]
|
2073
|
+
# Series to compare with.
|
2074
|
+
# @param strict [Boolean]
|
2075
|
+
# Require data types to match.
|
2076
|
+
# @param check_names [Boolean]
|
2077
|
+
# Require names to match.
|
2078
|
+
# @param null_equal [Boolean]
|
2079
|
+
# Consider null values as equal.
|
2080
|
+
#
|
2081
|
+
# @return [Boolean]
|
2082
|
+
#
|
2083
|
+
# @example
|
2084
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2085
|
+
# s2 = Polars::Series.new("b", [4, 5, 6])
|
2086
|
+
# s.equals(s)
|
2087
|
+
# # => true
|
2088
|
+
# s.equals(s2)
|
2089
|
+
# # => false
|
2090
|
+
def equals(other, strict: false, check_names: false, null_equal: false)
|
2091
|
+
_s.equals(other._s, strict, check_names, null_equal)
|
2092
|
+
end
|
2093
|
+
alias_method :series_equal, :equals
|
2094
|
+
|
2095
|
+
# Return the number of elements in the Series.
|
2096
|
+
#
|
2097
|
+
# @return [Integer]
|
2098
|
+
#
|
2099
|
+
# @example
|
2100
|
+
# s = Polars::Series.new("a", [1, 2, nil])
|
2101
|
+
# s.count
|
2102
|
+
# # => 2
|
2103
|
+
def count
|
2104
|
+
len - null_count
|
2105
|
+
end
|
2106
|
+
|
2107
|
+
# Return the number of elements in the Series.
|
2108
|
+
#
|
2109
|
+
# @return [Integer]
|
2110
|
+
#
|
2111
|
+
# @example
|
2112
|
+
# s = Polars::Series.new("a", [1, 2, nil])
|
2113
|
+
# s.len
|
2114
|
+
# # => 3
|
2115
|
+
def len
|
2116
|
+
_s.len
|
2117
|
+
end
|
2118
|
+
alias_method :length, :len
|
2119
|
+
alias_method :size, :len
|
2120
|
+
|
2121
|
+
# Cast between data types.
|
2122
|
+
#
|
2123
|
+
# @param dtype [Symbol]
|
2124
|
+
# DataType to cast to
|
2125
|
+
# @param strict [Boolean]
|
2126
|
+
# Throw an error if a cast could not be done for instance due to an overflow
|
2127
|
+
#
|
2128
|
+
# @return [Series]
|
2129
|
+
#
|
2130
|
+
# @example
|
2131
|
+
# s = Polars::Series.new("a", [true, false, true])
|
2132
|
+
# s.cast(:u32)
|
2133
|
+
# # =>
|
2134
|
+
# # shape: (3,)
|
2135
|
+
# # Series: 'a' [u32]
|
2136
|
+
# # [
|
2137
|
+
# # 1
|
2138
|
+
# # 0
|
2139
|
+
# # 1
|
2140
|
+
# # ]
|
2141
|
+
def cast(dtype, strict: true)
|
2142
|
+
super
|
2143
|
+
end
|
2144
|
+
|
2145
|
+
# Cast to physical representation of the logical dtype.
|
2146
|
+
#
|
2147
|
+
# - `:date` -> `:i32`
|
2148
|
+
# - `:datetime` -> `:i64`
|
2149
|
+
# - `:time` -> `:i64`
|
2150
|
+
# - `:duration` -> `:i64`
|
2151
|
+
# - `:cat` -> `:u32`
|
2152
|
+
# - other data types will be left unchanged.
|
2153
|
+
#
|
2154
|
+
# @return [Series]
|
2155
|
+
#
|
2156
|
+
# @example
|
2157
|
+
# s = Polars::Series.new("values", ["a", nil, "x", "a"])
|
2158
|
+
# s.cast(:cat).to_physical
|
2159
|
+
# # =>
|
2160
|
+
# # shape: (4,)
|
2161
|
+
# # Series: 'values' [u32]
|
2162
|
+
# # [
|
2163
|
+
# # 0
|
2164
|
+
# # null
|
2165
|
+
# # 1
|
2166
|
+
# # 0
|
2167
|
+
# # ]
|
2168
|
+
def to_physical
|
2169
|
+
super
|
2170
|
+
end
|
2171
|
+
|
2172
|
+
# Convert this Series to a Ruby Array. This operation clones data.
|
2173
|
+
#
|
2174
|
+
# @return [Array]
|
2175
|
+
#
|
2176
|
+
# @example
|
2177
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2178
|
+
# s.to_a
|
2179
|
+
# # => [1, 2, 3]
|
2180
|
+
def to_a
|
2181
|
+
_s.to_a
|
2182
|
+
end
|
2183
|
+
|
2184
|
+
# Create a single chunk of memory for this Series.
|
2185
|
+
#
|
2186
|
+
# @param in_place [Boolean]
|
2187
|
+
# In place or not.
|
2188
|
+
#
|
2189
|
+
# @return [Series]
|
2190
|
+
def rechunk(in_place: false)
|
2191
|
+
opt_s = _s.rechunk(in_place)
|
2192
|
+
in_place ? self : Utils.wrap_s(opt_s)
|
2193
|
+
end
|
2194
|
+
|
2195
|
+
# Return Series in reverse order.
|
2196
|
+
#
|
2197
|
+
# @return [Series]
|
2198
|
+
#
|
2199
|
+
# @example
|
2200
|
+
# s = Polars::Series.new("a", [1, 2, 3], dtype: :i8)
|
2201
|
+
# s.reverse
|
2202
|
+
# # =>
|
2203
|
+
# # shape: (3,)
|
2204
|
+
# # Series: 'a' [i8]
|
2205
|
+
# # [
|
2206
|
+
# # 3
|
2207
|
+
# # 2
|
2208
|
+
# # 1
|
2209
|
+
# # ]
|
2210
|
+
def reverse
|
2211
|
+
super
|
2212
|
+
end
|
2213
|
+
|
2214
|
+
# Check if this Series datatype is numeric.
|
2215
|
+
#
|
2216
|
+
# @return [Boolean]
|
2217
|
+
#
|
2218
|
+
# @example
|
2219
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2220
|
+
# s.is_numeric
|
2221
|
+
# # => true
|
2222
|
+
def is_numeric
|
2223
|
+
[Int8, Int16, Int32, Int64, UInt8, UInt16, UInt32, UInt64, Float32, Float64].include?(dtype)
|
2224
|
+
end
|
2225
|
+
alias_method :numeric?, :is_numeric
|
2226
|
+
|
2227
|
+
# Check if this Series datatype is datelike.
|
2228
|
+
#
|
2229
|
+
# @return [Boolean]
|
2230
|
+
#
|
2231
|
+
# @example
|
2232
|
+
# s = Polars::Series.new([Date.new(2021, 1, 1), Date.new(2021, 1, 2), Date.new(2021, 1, 3)])
|
2233
|
+
# s.is_datelike
|
2234
|
+
# # => true
|
2235
|
+
def is_datelike
|
2236
|
+
[Date, Time].include?(dtype) || dtype.is_a?(Datetime) || dtype.is_a?(Duration)
|
2237
|
+
end
|
2238
|
+
alias_method :datelike?, :is_datelike
|
2239
|
+
alias_method :is_temporal, :is_datelike
|
2240
|
+
alias_method :temporal?, :is_datelike
|
2241
|
+
|
2242
|
+
# Check if this Series has floating point numbers.
|
2243
|
+
#
|
2244
|
+
# @return [Boolean]
|
2245
|
+
#
|
2246
|
+
# @example
|
2247
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0])
|
2248
|
+
# s.is_float
|
2249
|
+
# # => true
|
2250
|
+
def is_float
|
2251
|
+
[Float32, Float64].include?(dtype)
|
2252
|
+
end
|
2253
|
+
alias_method :float?, :is_float
|
2254
|
+
|
2255
|
+
# Check if this Series is a Boolean.
|
2256
|
+
#
|
2257
|
+
# @return [Boolean]
|
2258
|
+
#
|
2259
|
+
# @example
|
2260
|
+
# s = Polars::Series.new("a", [true, false, true])
|
2261
|
+
# s.is_boolean
|
2262
|
+
# # => true
|
2263
|
+
def is_boolean
|
2264
|
+
dtype == Boolean
|
2265
|
+
end
|
2266
|
+
alias_method :boolean?, :is_boolean
|
2267
|
+
alias_method :is_bool, :is_boolean
|
2268
|
+
alias_method :bool?, :is_boolean
|
2269
|
+
|
2270
|
+
# Check if this Series datatype is a Utf8.
|
2271
|
+
#
|
2272
|
+
# @return [Boolean]
|
2273
|
+
#
|
2274
|
+
# @example
|
2275
|
+
# s = Polars::Series.new("x", ["a", "b", "c"])
|
2276
|
+
# s.is_utf8
|
2277
|
+
# # => true
|
2278
|
+
def is_utf8
|
2279
|
+
dtype == String
|
2280
|
+
end
|
2281
|
+
alias_method :utf8?, :is_utf8
|
2282
|
+
|
2283
|
+
# def view
|
2284
|
+
# end
|
2285
|
+
|
2286
|
+
# Convert this Series to a Numo array. This operation clones data but is completely safe.
|
2287
|
+
#
|
2288
|
+
# @return [Numo::NArray]
|
2289
|
+
#
|
2290
|
+
# @example
|
2291
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2292
|
+
# s.to_numo
|
2293
|
+
# # =>
|
2294
|
+
# # Numo::Int64#shape=[3]
|
2295
|
+
# # [1, 2, 3]
|
2296
|
+
def to_numo
|
2297
|
+
if !has_validity
|
2298
|
+
if is_datelike
|
2299
|
+
Numo::RObject.cast(to_a)
|
2300
|
+
elsif is_numeric
|
2301
|
+
# TODO make more efficient
|
2302
|
+
{
|
2303
|
+
UInt8 => Numo::UInt8,
|
2304
|
+
UInt16 => Numo::UInt16,
|
2305
|
+
UInt32 => Numo::UInt32,
|
2306
|
+
UInt64 => Numo::UInt64,
|
2307
|
+
Int8 => Numo::Int8,
|
2308
|
+
Int16 => Numo::Int16,
|
2309
|
+
Int32 => Numo::Int32,
|
2310
|
+
Int64 => Numo::Int64,
|
2311
|
+
Float32 => Numo::SFloat,
|
2312
|
+
Float64 => Numo::DFloat
|
2313
|
+
}.fetch(dtype.class).cast(to_a)
|
2314
|
+
elsif is_boolean
|
2315
|
+
Numo::Bit.cast(to_a)
|
2316
|
+
else
|
2317
|
+
_s.to_numo
|
2318
|
+
end
|
2319
|
+
elsif is_datelike
|
2320
|
+
Numo::RObject.cast(to_a)
|
2321
|
+
else
|
2322
|
+
_s.to_numo
|
2323
|
+
end
|
2324
|
+
end
|
2325
|
+
|
2326
|
+
# Set masked values.
|
2327
|
+
#
|
2328
|
+
# @param filter [Series]
|
2329
|
+
# Boolean mask.
|
2330
|
+
# @param value [Object]
|
2331
|
+
# Value with which to replace the masked values.
|
2332
|
+
#
|
2333
|
+
# @return [Series]
|
2334
|
+
#
|
2335
|
+
# @note
|
2336
|
+
# Use of this function is frequently an anti-pattern, as it can
|
2337
|
+
# block optimization (predicate pushdown, etc). Consider using
|
2338
|
+
# `Polars.when(predicate).then(value).otherwise(self)` instead.
|
2339
|
+
#
|
2340
|
+
# @example
|
2341
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2342
|
+
# s.set(s == 2, 10)
|
2343
|
+
# # =>
|
2344
|
+
# # shape: (3,)
|
2345
|
+
# # Series: 'a' [i64]
|
2346
|
+
# # [
|
2347
|
+
# # 1
|
2348
|
+
# # 10
|
2349
|
+
# # 3
|
2350
|
+
# # ]
|
2351
|
+
def set(filter, value)
|
2352
|
+
Utils.wrap_s(_s.send("set_with_mask_#{DTYPE_TO_FFINAME.fetch(dtype.class)}", filter._s, value))
|
2353
|
+
end
|
2354
|
+
|
2355
|
+
# Set values at the index locations.
|
2356
|
+
#
|
2357
|
+
# @param idx [Object]
|
2358
|
+
# Integers representing the index locations.
|
2359
|
+
# @param value [Object]
|
2360
|
+
# Replacement values.
|
2361
|
+
#
|
2362
|
+
# @return [Series]
|
2363
|
+
#
|
2364
|
+
# @example
|
2365
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2366
|
+
# s.set_at_idx(1, 10)
|
2367
|
+
# # =>
|
2368
|
+
# # shape: (3,)
|
2369
|
+
# # Series: 'a' [i64]
|
2370
|
+
# # [
|
2371
|
+
# # 1
|
2372
|
+
# # 10
|
2373
|
+
# # 3
|
2374
|
+
# # ]
|
2375
|
+
def scatter(idx, value)
|
2376
|
+
if idx.is_a?(Integer)
|
2377
|
+
idx = [idx]
|
2378
|
+
end
|
2379
|
+
if idx.length == 0
|
2380
|
+
return self
|
2381
|
+
end
|
2382
|
+
|
2383
|
+
idx = Series.new("", idx)
|
2384
|
+
if value.is_a?(Integer) || value.is_a?(Float) || Utils.bool?(value) || value.is_a?(::String) || value.nil?
|
2385
|
+
value = Series.new("", [value])
|
2386
|
+
|
2387
|
+
# if we need to set more than a single value, we extend it
|
2388
|
+
if idx.length > 0
|
2389
|
+
value = value.extend_constant(value[0], idx.length - 1)
|
2390
|
+
end
|
2391
|
+
elsif !value.is_a?(Series)
|
2392
|
+
value = Series.new("", value)
|
2393
|
+
end
|
2394
|
+
_s.scatter(idx._s, value._s)
|
2395
|
+
self
|
2396
|
+
end
|
2397
|
+
alias_method :set_at_idx, :scatter
|
2398
|
+
|
2399
|
+
# Create an empty copy of the current Series.
|
2400
|
+
#
|
2401
|
+
# The copy has identical name/dtype but no data.
|
2402
|
+
#
|
2403
|
+
# @return [Series]
|
2404
|
+
#
|
2405
|
+
# @example
|
2406
|
+
# s = Polars::Series.new("a", [nil, true, false])
|
2407
|
+
# s.cleared
|
2408
|
+
# # =>
|
2409
|
+
# # shape: (0,)
|
2410
|
+
# # Series: 'a' [bool]
|
2411
|
+
# # [
|
2412
|
+
# # ]
|
2413
|
+
def cleared
|
2414
|
+
len > 0 ? limit(0) : clone
|
2415
|
+
end
|
2416
|
+
|
2417
|
+
# clone handled by initialize_copy
|
2418
|
+
|
2419
|
+
# Fill floating point NaN value with a fill value.
|
2420
|
+
#
|
2421
|
+
# @param fill_value [Object]
|
2422
|
+
# Value used to fill nan values.
|
2423
|
+
#
|
2424
|
+
# @return [Series]
|
2425
|
+
#
|
2426
|
+
# @example
|
2427
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, Float::NAN])
|
2428
|
+
# s.fill_nan(0)
|
2429
|
+
# # =>
|
2430
|
+
# # shape: (4,)
|
2431
|
+
# # Series: 'a' [f64]
|
2432
|
+
# # [
|
2433
|
+
# # 1.0
|
2434
|
+
# # 2.0
|
2435
|
+
# # 3.0
|
2436
|
+
# # 0.0
|
2437
|
+
# # ]
|
2438
|
+
def fill_nan(fill_value)
|
2439
|
+
super
|
2440
|
+
end
|
2441
|
+
|
2442
|
+
# Fill null values using the specified value or strategy.
|
2443
|
+
#
|
2444
|
+
# @param value [Object]
|
2445
|
+
# Value used to fill null values.
|
2446
|
+
# @param strategy [nil, "forward", "backward", "min", "max", "mean", "zero", "one"]
|
2447
|
+
# Strategy used to fill null values.
|
2448
|
+
# @param limit
|
2449
|
+
# Number of consecutive null values to fill when using the "forward" or
|
2450
|
+
# "backward" strategy.
|
2451
|
+
#
|
2452
|
+
# @return [Series]
|
2453
|
+
#
|
2454
|
+
# @example
|
2455
|
+
# s = Polars::Series.new("a", [1, 2, 3, nil])
|
2456
|
+
# s.fill_null(strategy: "forward")
|
2457
|
+
# # =>
|
2458
|
+
# # shape: (4,)
|
2459
|
+
# # Series: 'a' [i64]
|
2460
|
+
# # [
|
2461
|
+
# # 1
|
2462
|
+
# # 2
|
2463
|
+
# # 3
|
2464
|
+
# # 3
|
2465
|
+
# # ]
|
2466
|
+
#
|
2467
|
+
# @example
|
2468
|
+
# s.fill_null(strategy: "min")
|
2469
|
+
# # =>
|
2470
|
+
# # shape: (4,)
|
2471
|
+
# # Series: 'a' [i64]
|
2472
|
+
# # [
|
2473
|
+
# # 1
|
2474
|
+
# # 2
|
2475
|
+
# # 3
|
2476
|
+
# # 1
|
2477
|
+
# # ]
|
2478
|
+
#
|
2479
|
+
# @example
|
2480
|
+
# s = Polars::Series.new("b", ["x", nil, "z"])
|
2481
|
+
# s.fill_null(Polars.lit(""))
|
2482
|
+
# # =>
|
2483
|
+
# # shape: (3,)
|
2484
|
+
# # Series: 'b' [str]
|
2485
|
+
# # [
|
2486
|
+
# # "x"
|
2487
|
+
# # ""
|
2488
|
+
# # "z"
|
2489
|
+
# # ]
|
2490
|
+
def fill_null(value = nil, strategy: nil, limit: nil)
|
2491
|
+
super
|
2492
|
+
end
|
2493
|
+
|
2494
|
+
# Rounds down to the nearest integer value.
|
2495
|
+
#
|
2496
|
+
# Only works on floating point Series.
|
2497
|
+
#
|
2498
|
+
# @return [Series]
|
2499
|
+
#
|
2500
|
+
# @example
|
2501
|
+
# s = Polars::Series.new("a", [1.12345, 2.56789, 3.901234])
|
2502
|
+
# s.floor
|
2503
|
+
# # =>
|
2504
|
+
# # shape: (3,)
|
2505
|
+
# # Series: 'a' [f64]
|
2506
|
+
# # [
|
2507
|
+
# # 1.0
|
2508
|
+
# # 2.0
|
2509
|
+
# # 3.0
|
2510
|
+
# # ]
|
2511
|
+
def floor
|
2512
|
+
Utils.wrap_s(_s.floor)
|
2513
|
+
end
|
2514
|
+
|
2515
|
+
# Rounds up to the nearest integer value.
|
2516
|
+
#
|
2517
|
+
# Only works on floating point Series.
|
2518
|
+
#
|
2519
|
+
# @return [Series]
|
2520
|
+
#
|
2521
|
+
# @example
|
2522
|
+
# s = Polars::Series.new("a", [1.12345, 2.56789, 3.901234])
|
2523
|
+
# s.ceil
|
2524
|
+
# # =>
|
2525
|
+
# # shape: (3,)
|
2526
|
+
# # Series: 'a' [f64]
|
2527
|
+
# # [
|
2528
|
+
# # 2.0
|
2529
|
+
# # 3.0
|
2530
|
+
# # 4.0
|
2531
|
+
# # ]
|
2532
|
+
def ceil
|
2533
|
+
super
|
2534
|
+
end
|
2535
|
+
|
2536
|
+
# Round underlying floating point data by `decimals` digits.
|
2537
|
+
#
|
2538
|
+
# @param decimals [Integer]
|
2539
|
+
# number of decimals to round by.
|
2540
|
+
#
|
2541
|
+
# @return [Series]
|
2542
|
+
#
|
2543
|
+
# @example
|
2544
|
+
# s = Polars::Series.new("a", [1.12345, 2.56789, 3.901234])
|
2545
|
+
# s.round(2)
|
2546
|
+
# # =>
|
2547
|
+
# # shape: (3,)
|
2548
|
+
# # Series: 'a' [f64]
|
2549
|
+
# # [
|
2550
|
+
# # 1.12
|
2551
|
+
# # 2.57
|
2552
|
+
# # 3.9
|
2553
|
+
# # ]
|
2554
|
+
def round(decimals = 0)
|
2555
|
+
super
|
2556
|
+
end
|
2557
|
+
|
2558
|
+
# Compute the dot/inner product between two Series.
|
2559
|
+
#
|
2560
|
+
# @param other [Object]
|
2561
|
+
# Series (or array) to compute dot product with.
|
2562
|
+
#
|
2563
|
+
# @return [Numeric]
|
2564
|
+
#
|
2565
|
+
# @example
|
2566
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2567
|
+
# s2 = Polars::Series.new("b", [4.0, 5.0, 6.0])
|
2568
|
+
# s.dot(s2)
|
2569
|
+
# # => 32.0
|
2570
|
+
def dot(other)
|
2571
|
+
if !other.is_a?(Series)
|
2572
|
+
other = Series.new(other)
|
2573
|
+
end
|
2574
|
+
if len != other.len
|
2575
|
+
n, m = len, other.len
|
2576
|
+
raise ArgumentError, "Series length mismatch: expected #{n}, found #{m}"
|
2577
|
+
end
|
2578
|
+
_s.dot(other._s)
|
2579
|
+
end
|
2580
|
+
|
2581
|
+
# Compute the most occurring value(s).
|
2582
|
+
#
|
2583
|
+
# Can return multiple Values.
|
2584
|
+
#
|
2585
|
+
# @return [Series]
|
2586
|
+
#
|
2587
|
+
# @example
|
2588
|
+
# s = Polars::Series.new("a", [1, 2, 2, 3])
|
2589
|
+
# s.mode
|
2590
|
+
# # =>
|
2591
|
+
# # shape: (1,)
|
2592
|
+
# # Series: 'a' [i64]
|
2593
|
+
# # [
|
2594
|
+
# # 2
|
2595
|
+
# # ]
|
2596
|
+
def mode
|
2597
|
+
super
|
2598
|
+
end
|
2599
|
+
|
2600
|
+
# Compute the element-wise indication of the sign.
|
2601
|
+
#
|
2602
|
+
# @return [Series]
|
2603
|
+
#
|
2604
|
+
# @example
|
2605
|
+
# s = Polars::Series.new("a", [-9.0, -0.0, 0.0, 4.0, nil])
|
2606
|
+
# s.sign
|
2607
|
+
# # =>
|
2608
|
+
# # shape: (5,)
|
2609
|
+
# # Series: 'a' [i64]
|
2610
|
+
# # [
|
2611
|
+
# # -1
|
2612
|
+
# # 0
|
2613
|
+
# # 0
|
2614
|
+
# # 1
|
2615
|
+
# # null
|
2616
|
+
# # ]
|
2617
|
+
def sign
|
2618
|
+
super
|
2619
|
+
end
|
2620
|
+
|
2621
|
+
# Compute the element-wise value for the sine.
|
2622
|
+
#
|
2623
|
+
# @return [Series]
|
2624
|
+
#
|
2625
|
+
# @example
|
2626
|
+
# s = Polars::Series.new("a", [0.0, Math::PI / 2.0, Math::PI])
|
2627
|
+
# s.sin
|
2628
|
+
# # =>
|
2629
|
+
# # shape: (3,)
|
2630
|
+
# # Series: 'a' [f64]
|
2631
|
+
# # [
|
2632
|
+
# # 0.0
|
2633
|
+
# # 1.0
|
2634
|
+
# # 1.2246e-16
|
2635
|
+
# # ]
|
2636
|
+
def sin
|
2637
|
+
super
|
2638
|
+
end
|
2639
|
+
|
2640
|
+
# Compute the element-wise value for the cosine.
|
2641
|
+
#
|
2642
|
+
# @return [Series]
|
2643
|
+
#
|
2644
|
+
# @example
|
2645
|
+
# s = Polars::Series.new("a", [0.0, Math::PI / 2.0, Math::PI])
|
2646
|
+
# s.cos
|
2647
|
+
# # =>
|
2648
|
+
# # shape: (3,)
|
2649
|
+
# # Series: 'a' [f64]
|
2650
|
+
# # [
|
2651
|
+
# # 1.0
|
2652
|
+
# # 6.1232e-17
|
2653
|
+
# # -1.0
|
2654
|
+
# # ]
|
2655
|
+
def cos
|
2656
|
+
super
|
2657
|
+
end
|
2658
|
+
|
2659
|
+
# Compute the element-wise value for the tangent.
|
2660
|
+
#
|
2661
|
+
# @return [Series]
|
2662
|
+
#
|
2663
|
+
# @example
|
2664
|
+
# s = Polars::Series.new("a", [0.0, Math::PI / 2.0, Math::PI])
|
2665
|
+
# s.tan
|
2666
|
+
# # =>
|
2667
|
+
# # shape: (3,)
|
2668
|
+
# # Series: 'a' [f64]
|
2669
|
+
# # [
|
2670
|
+
# # 0.0
|
2671
|
+
# # 1.6331e16
|
2672
|
+
# # -1.2246e-16
|
2673
|
+
# # ]
|
2674
|
+
def tan
|
2675
|
+
super
|
2676
|
+
end
|
2677
|
+
|
2678
|
+
# Compute the element-wise value for the inverse sine.
|
2679
|
+
#
|
2680
|
+
# @return [Series]
|
2681
|
+
#
|
2682
|
+
# @example
|
2683
|
+
# s = Polars::Series.new("a", [1.0, 0.0, -1.0])
|
2684
|
+
# s.arcsin
|
2685
|
+
# # =>
|
2686
|
+
# # shape: (3,)
|
2687
|
+
# # Series: 'a' [f64]
|
2688
|
+
# # [
|
2689
|
+
# # 1.570796
|
2690
|
+
# # 0.0
|
2691
|
+
# # -1.570796
|
2692
|
+
# # ]
|
2693
|
+
def arcsin
|
2694
|
+
super
|
2695
|
+
end
|
2696
|
+
alias_method :asin, :arcsin
|
2697
|
+
|
2698
|
+
# Compute the element-wise value for the inverse cosine.
|
2699
|
+
#
|
2700
|
+
# @return [Series]
|
2701
|
+
#
|
2702
|
+
# @example
|
2703
|
+
# s = Polars::Series.new("a", [1.0, 0.0, -1.0])
|
2704
|
+
# s.arccos
|
2705
|
+
# # =>
|
2706
|
+
# # shape: (3,)
|
2707
|
+
# # Series: 'a' [f64]
|
2708
|
+
# # [
|
2709
|
+
# # 0.0
|
2710
|
+
# # 1.570796
|
2711
|
+
# # 3.141593
|
2712
|
+
# # ]
|
2713
|
+
def arccos
|
2714
|
+
super
|
2715
|
+
end
|
2716
|
+
alias_method :acos, :arccos
|
2717
|
+
|
2718
|
+
# Compute the element-wise value for the inverse tangent.
|
2719
|
+
#
|
2720
|
+
# @return [Series]
|
2721
|
+
#
|
2722
|
+
# @example
|
2723
|
+
# s = Polars::Series.new("a", [1.0, 0.0, -1.0])
|
2724
|
+
# s.arctan
|
2725
|
+
# # =>
|
2726
|
+
# # shape: (3,)
|
2727
|
+
# # Series: 'a' [f64]
|
2728
|
+
# # [
|
2729
|
+
# # 0.785398
|
2730
|
+
# # 0.0
|
2731
|
+
# # -0.785398
|
2732
|
+
# # ]
|
2733
|
+
def arctan
|
2734
|
+
super
|
2735
|
+
end
|
2736
|
+
alias_method :atan, :arctan
|
2737
|
+
|
2738
|
+
# Compute the element-wise value for the inverse hyperbolic sine.
|
2739
|
+
#
|
2740
|
+
# @return [Series]
|
2741
|
+
#
|
2742
|
+
# @example
|
2743
|
+
# s = Polars::Series.new("a", [1.0, 0.0, -1.0])
|
2744
|
+
# s.arcsinh
|
2745
|
+
# # =>
|
2746
|
+
# # shape: (3,)
|
2747
|
+
# # Series: 'a' [f64]
|
2748
|
+
# # [
|
2749
|
+
# # 0.881374
|
2750
|
+
# # 0.0
|
2751
|
+
# # -0.881374
|
2752
|
+
# # ]
|
2753
|
+
def arcsinh
|
2754
|
+
super
|
2755
|
+
end
|
2756
|
+
alias_method :asinh, :arcsinh
|
2757
|
+
|
2758
|
+
# Compute the element-wise value for the inverse hyperbolic cosine.
|
2759
|
+
#
|
2760
|
+
# @return [Series]
|
2761
|
+
#
|
2762
|
+
# @example
|
2763
|
+
# s = Polars::Series.new("a", [5.0, 1.0, 0.0, -1.0])
|
2764
|
+
# s.arccosh
|
2765
|
+
# # =>
|
2766
|
+
# # shape: (4,)
|
2767
|
+
# # Series: 'a' [f64]
|
2768
|
+
# # [
|
2769
|
+
# # 2.292432
|
2770
|
+
# # 0.0
|
2771
|
+
# # NaN
|
2772
|
+
# # NaN
|
2773
|
+
# # ]
|
2774
|
+
def arccosh
|
2775
|
+
super
|
2776
|
+
end
|
2777
|
+
alias_method :acosh, :arccosh
|
2778
|
+
|
2779
|
+
# Compute the element-wise value for the inverse hyperbolic tangent.
|
2780
|
+
#
|
2781
|
+
# @return [Series]
|
2782
|
+
#
|
2783
|
+
# @example
|
2784
|
+
# s = Polars::Series.new("a", [2.0, 1.0, 0.5, 0.0, -0.5, -1.0, -1.1])
|
2785
|
+
# s.arctanh
|
2786
|
+
# # =>
|
2787
|
+
# # shape: (7,)
|
2788
|
+
# # Series: 'a' [f64]
|
2789
|
+
# # [
|
2790
|
+
# # NaN
|
2791
|
+
# # inf
|
2792
|
+
# # 0.549306
|
2793
|
+
# # 0.0
|
2794
|
+
# # -0.549306
|
2795
|
+
# # -inf
|
2796
|
+
# # NaN
|
2797
|
+
# # ]
|
2798
|
+
def arctanh
|
2799
|
+
super
|
2800
|
+
end
|
2801
|
+
alias_method :atanh, :arctanh
|
2802
|
+
|
2803
|
+
# Compute the element-wise value for the hyperbolic sine.
|
2804
|
+
#
|
2805
|
+
# @return [Series]
|
2806
|
+
#
|
2807
|
+
# @example
|
2808
|
+
# s = Polars::Series.new("a", [1.0, 0.0, -1.0])
|
2809
|
+
# s.sinh
|
2810
|
+
# # =>
|
2811
|
+
# # shape: (3,)
|
2812
|
+
# # Series: 'a' [f64]
|
2813
|
+
# # [
|
2814
|
+
# # 1.175201
|
2815
|
+
# # 0.0
|
2816
|
+
# # -1.175201
|
2817
|
+
# # ]
|
2818
|
+
def sinh
|
2819
|
+
super
|
2820
|
+
end
|
2821
|
+
|
2822
|
+
# Compute the element-wise value for the hyperbolic cosine.
|
2823
|
+
#
|
2824
|
+
# @return [Series]
|
2825
|
+
#
|
2826
|
+
# @example
|
2827
|
+
# s = Polars::Series.new("a", [1.0, 0.0, -1.0])
|
2828
|
+
# s.cosh
|
2829
|
+
# # =>
|
2830
|
+
# # shape: (3,)
|
2831
|
+
# # Series: 'a' [f64]
|
2832
|
+
# # [
|
2833
|
+
# # 1.543081
|
2834
|
+
# # 1.0
|
2835
|
+
# # 1.543081
|
2836
|
+
# # ]
|
2837
|
+
def cosh
|
2838
|
+
super
|
2839
|
+
end
|
2840
|
+
|
2841
|
+
# Compute the element-wise value for the hyperbolic tangent.
|
2842
|
+
#
|
2843
|
+
# @return [Series]
|
2844
|
+
#
|
2845
|
+
# @example
|
2846
|
+
# s = Polars::Series.new("a", [1.0, 0.0, -1.0])
|
2847
|
+
# s.tanh
|
2848
|
+
# # =>
|
2849
|
+
# # shape: (3,)
|
2850
|
+
# # Series: 'a' [f64]
|
2851
|
+
# # [
|
2852
|
+
# # 0.761594
|
2853
|
+
# # 0.0
|
2854
|
+
# # -0.761594
|
2855
|
+
# # ]
|
2856
|
+
def tanh
|
2857
|
+
super
|
2858
|
+
end
|
2859
|
+
|
2860
|
+
# Apply a custom/user-defined function (UDF) over elements in this Series and
|
2861
|
+
# return a new Series.
|
2862
|
+
#
|
2863
|
+
# If the function returns another datatype, the return_dtype arg should be set,
|
2864
|
+
# otherwise the method will fail.
|
2865
|
+
#
|
2866
|
+
# @param return_dtype [Symbol]
|
2867
|
+
# Output datatype. If none is given, the same datatype as this Series will be
|
2868
|
+
# used.
|
2869
|
+
# @param skip_nulls [Boolean]
|
2870
|
+
# Nulls will be skipped and not passed to the Ruby function.
|
2871
|
+
# This is faster because Ruby can be skipped and because we call
|
2872
|
+
# more specialized functions.
|
2873
|
+
#
|
2874
|
+
# @return [Series]
|
2875
|
+
#
|
2876
|
+
# @example
|
2877
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2878
|
+
# s.map_elements { |x| x + 10 }
|
2879
|
+
# # =>
|
2880
|
+
# # shape: (3,)
|
2881
|
+
# # Series: 'a' [i64]
|
2882
|
+
# # [
|
2883
|
+
# # 11
|
2884
|
+
# # 12
|
2885
|
+
# # 13
|
2886
|
+
# # ]
|
2887
|
+
def map_elements(return_dtype: nil, skip_nulls: true, &func)
|
2888
|
+
if return_dtype.nil?
|
2889
|
+
pl_return_dtype = nil
|
2890
|
+
else
|
2891
|
+
pl_return_dtype = Utils.rb_type_to_dtype(return_dtype)
|
2892
|
+
end
|
2893
|
+
Utils.wrap_s(_s.apply_lambda(func, pl_return_dtype, skip_nulls))
|
2894
|
+
end
|
2895
|
+
alias_method :map, :map_elements
|
2896
|
+
alias_method :apply, :map_elements
|
2897
|
+
|
2898
|
+
# Shift the values by a given period.
|
2899
|
+
#
|
2900
|
+
# @param periods [Integer]
|
2901
|
+
# Number of places to shift (may be negative).
|
2902
|
+
#
|
2903
|
+
# @return [Series]
|
2904
|
+
#
|
2905
|
+
# @example
|
2906
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
2907
|
+
# s.shift(1)
|
2908
|
+
# # =>
|
2909
|
+
# # shape: (3,)
|
2910
|
+
# # Series: 'a' [i64]
|
2911
|
+
# # [
|
2912
|
+
# # null
|
2913
|
+
# # 1
|
2914
|
+
# # 2
|
2915
|
+
# # ]
|
2916
|
+
#
|
2917
|
+
# @example
|
2918
|
+
# s.shift(-1)
|
2919
|
+
# # =>
|
2920
|
+
# # shape: (3,)
|
2921
|
+
# # Series: 'a' [i64]
|
2922
|
+
# # [
|
2923
|
+
# # 2
|
2924
|
+
# # 3
|
2925
|
+
# # null
|
2926
|
+
# # ]
|
2927
|
+
def shift(periods = 1)
|
2928
|
+
super
|
2929
|
+
end
|
2930
|
+
|
2931
|
+
# Shift the values by a given period and fill the resulting null values.
|
2932
|
+
#
|
2933
|
+
# @param periods [Integer]
|
2934
|
+
# Number of places to shift (may be negative).
|
2935
|
+
# @param fill_value [Object]
|
2936
|
+
# Fill None values with the result of this expression.
|
2937
|
+
#
|
2938
|
+
# @return [Series]
|
2939
|
+
def shift_and_fill(periods, fill_value)
|
2940
|
+
super
|
2941
|
+
end
|
2942
|
+
|
2943
|
+
# Take values from self or other based on the given mask.
|
2944
|
+
#
|
2945
|
+
# Where mask evaluates true, take values from self. Where mask evaluates false,
|
2946
|
+
# take values from other.
|
2947
|
+
#
|
2948
|
+
# @param mask [Series]
|
2949
|
+
# Boolean Series.
|
2950
|
+
# @param other [Series]
|
2951
|
+
# Series of same type.
|
2952
|
+
#
|
2953
|
+
# @return [Series]
|
2954
|
+
#
|
2955
|
+
# @example
|
2956
|
+
# s1 = Polars::Series.new([1, 2, 3, 4, 5])
|
2957
|
+
# s2 = Polars::Series.new([5, 4, 3, 2, 1])
|
2958
|
+
# s1.zip_with(s1 < s2, s2)
|
2959
|
+
# # =>
|
2960
|
+
# # shape: (5,)
|
2961
|
+
# # Series: '' [i64]
|
2962
|
+
# # [
|
2963
|
+
# # 1
|
2964
|
+
# # 2
|
2965
|
+
# # 3
|
2966
|
+
# # 2
|
2967
|
+
# # 1
|
2968
|
+
# # ]
|
2969
|
+
#
|
2970
|
+
# @example
|
2971
|
+
# mask = Polars::Series.new([true, false, true, false, true])
|
2972
|
+
# s1.zip_with(mask, s2)
|
2973
|
+
# # =>
|
2974
|
+
# # shape: (5,)
|
2975
|
+
# # Series: '' [i64]
|
2976
|
+
# # [
|
2977
|
+
# # 1
|
2978
|
+
# # 4
|
2979
|
+
# # 3
|
2980
|
+
# # 2
|
2981
|
+
# # 5
|
2982
|
+
# # ]
|
2983
|
+
def zip_with(mask, other)
|
2984
|
+
Utils.wrap_s(_s.zip_with(mask._s, other._s))
|
2985
|
+
end
|
2986
|
+
|
2987
|
+
# Apply a rolling min (moving min) over the values in this array.
|
2988
|
+
#
|
2989
|
+
# A window of length `window_size` will traverse the array. The values that fill
|
2990
|
+
# this window will (optionally) be multiplied with the weights given by the
|
2991
|
+
# `weight` vector. The resulting values will be aggregated to their sum.
|
2992
|
+
#
|
2993
|
+
# @param window_size [Integer]
|
2994
|
+
# The length of the window.
|
2995
|
+
# @param weights [Array]
|
2996
|
+
# An optional slice with the same length as the window that will be multiplied
|
2997
|
+
# elementwise with the values in the window.
|
2998
|
+
# @param min_periods [Integer]
|
2999
|
+
# The number of values in the window that should be non-null before computing
|
3000
|
+
# a result. If None, it will be set equal to window size.
|
3001
|
+
# @param center [Boolean]
|
3002
|
+
# Set the labels at the center of the window
|
3003
|
+
#
|
3004
|
+
# @return [Series]
|
3005
|
+
#
|
3006
|
+
# @example
|
3007
|
+
# s = Polars::Series.new("a", [100, 200, 300, 400, 500])
|
3008
|
+
# s.rolling_min(3)
|
3009
|
+
# # =>
|
3010
|
+
# # shape: (5,)
|
3011
|
+
# # Series: 'a' [i64]
|
3012
|
+
# # [
|
3013
|
+
# # null
|
3014
|
+
# # null
|
3015
|
+
# # 100
|
3016
|
+
# # 200
|
3017
|
+
# # 300
|
3018
|
+
# # ]
|
3019
|
+
def rolling_min(
|
3020
|
+
window_size,
|
3021
|
+
weights: nil,
|
3022
|
+
min_periods: nil,
|
3023
|
+
center: false
|
3024
|
+
)
|
3025
|
+
super
|
3026
|
+
end
|
3027
|
+
|
3028
|
+
# Apply a rolling max (moving max) over the values in this array.
|
3029
|
+
#
|
3030
|
+
# A window of length `window_size` will traverse the array. The values that fill
|
3031
|
+
# this window will (optionally) be multiplied with the weights given by the
|
3032
|
+
# `weight` vector. The resulting values will be aggregated to their sum.
|
3033
|
+
#
|
3034
|
+
# @param window_size [Integer]
|
3035
|
+
# The length of the window.
|
3036
|
+
# @param weights [Array]
|
3037
|
+
# An optional slice with the same length as the window that will be multiplied
|
3038
|
+
# elementwise with the values in the window.
|
3039
|
+
# @param min_periods [Integer]
|
3040
|
+
# The number of values in the window that should be non-null before computing
|
3041
|
+
# a result. If None, it will be set equal to window size.
|
3042
|
+
# @param center [Boolean]
|
3043
|
+
# Set the labels at the center of the window
|
3044
|
+
#
|
3045
|
+
# @return [Series]
|
3046
|
+
#
|
3047
|
+
# @example
|
3048
|
+
# s = Polars::Series.new("a", [100, 200, 300, 400, 500])
|
3049
|
+
# s.rolling_max(2)
|
3050
|
+
# # =>
|
3051
|
+
# # shape: (5,)
|
3052
|
+
# # Series: 'a' [i64]
|
3053
|
+
# # [
|
3054
|
+
# # null
|
3055
|
+
# # 200
|
3056
|
+
# # 300
|
3057
|
+
# # 400
|
3058
|
+
# # 500
|
3059
|
+
# # ]
|
3060
|
+
def rolling_max(
|
3061
|
+
window_size,
|
3062
|
+
weights: nil,
|
3063
|
+
min_periods: nil,
|
3064
|
+
center: false
|
3065
|
+
)
|
3066
|
+
super
|
3067
|
+
end
|
3068
|
+
|
3069
|
+
# Apply a rolling mean (moving mean) over the values in this array.
|
3070
|
+
#
|
3071
|
+
# A window of length `window_size` will traverse the array. The values that fill
|
3072
|
+
# this window will (optionally) be multiplied with the weights given by the
|
3073
|
+
# `weight` vector. The resulting values will be aggregated to their sum.
|
3074
|
+
#
|
3075
|
+
# @param window_size [Integer]
|
3076
|
+
# The length of the window.
|
3077
|
+
# @param weights [Array]
|
3078
|
+
# An optional slice with the same length as the window that will be multiplied
|
3079
|
+
# elementwise with the values in the window.
|
3080
|
+
# @param min_periods [Integer]
|
3081
|
+
# The number of values in the window that should be non-null before computing
|
3082
|
+
# a result. If None, it will be set equal to window size.
|
3083
|
+
# @param center [Boolean]
|
3084
|
+
# Set the labels at the center of the window
|
3085
|
+
#
|
3086
|
+
# @return [Series]
|
3087
|
+
#
|
3088
|
+
# @example
|
3089
|
+
# s = Polars::Series.new("a", [100, 200, 300, 400, 500])
|
3090
|
+
# s.rolling_mean(2)
|
3091
|
+
# # =>
|
3092
|
+
# # shape: (5,)
|
3093
|
+
# # Series: 'a' [f64]
|
3094
|
+
# # [
|
3095
|
+
# # null
|
3096
|
+
# # 150.0
|
3097
|
+
# # 250.0
|
3098
|
+
# # 350.0
|
3099
|
+
# # 450.0
|
3100
|
+
# # ]
|
3101
|
+
def rolling_mean(
|
3102
|
+
window_size,
|
3103
|
+
weights: nil,
|
3104
|
+
min_periods: nil,
|
3105
|
+
center: false
|
3106
|
+
)
|
3107
|
+
super
|
3108
|
+
end
|
3109
|
+
|
3110
|
+
# Apply a rolling sum (moving sum) over the values in this array.
|
3111
|
+
#
|
3112
|
+
# A window of length `window_size` will traverse the array. The values that fill
|
3113
|
+
# this window will (optionally) be multiplied with the weights given by the
|
3114
|
+
# `weight` vector. The resulting values will be aggregated to their sum.
|
3115
|
+
#
|
3116
|
+
# @param window_size [Integer]
|
3117
|
+
# The length of the window.
|
3118
|
+
# @param weights [Array]
|
3119
|
+
# An optional slice with the same length as the window that will be multiplied
|
3120
|
+
# elementwise with the values in the window.
|
3121
|
+
# @param min_periods [Integer]
|
3122
|
+
# The number of values in the window that should be non-null before computing
|
3123
|
+
# a result. If None, it will be set equal to window size.
|
3124
|
+
# @param center [Boolean]
|
3125
|
+
# Set the labels at the center of the window
|
3126
|
+
#
|
3127
|
+
# @return [Series]
|
3128
|
+
#
|
3129
|
+
# @example
|
3130
|
+
# s = Polars::Series.new("a", [1, 2, 3, 4, 5])
|
3131
|
+
# s.rolling_sum(2)
|
3132
|
+
# # =>
|
3133
|
+
# # shape: (5,)
|
3134
|
+
# # Series: 'a' [i64]
|
3135
|
+
# # [
|
3136
|
+
# # null
|
3137
|
+
# # 3
|
3138
|
+
# # 5
|
3139
|
+
# # 7
|
3140
|
+
# # 9
|
3141
|
+
# # ]
|
3142
|
+
def rolling_sum(
|
3143
|
+
window_size,
|
3144
|
+
weights: nil,
|
3145
|
+
min_periods: nil,
|
3146
|
+
center: false
|
3147
|
+
)
|
3148
|
+
super
|
3149
|
+
end
|
3150
|
+
|
3151
|
+
# Compute a rolling std dev.
|
3152
|
+
#
|
3153
|
+
# A window of length `window_size` will traverse the array. The values that fill
|
3154
|
+
# this window will (optionally) be multiplied with the weights given by the
|
3155
|
+
# `weight` vector. The resulting values will be aggregated to their sum.
|
3156
|
+
#
|
3157
|
+
# @param window_size [Integer]
|
3158
|
+
# The length of the window.
|
3159
|
+
# @param weights [Array]
|
3160
|
+
# An optional slice with the same length as the window that will be multiplied
|
3161
|
+
# elementwise with the values in the window.
|
3162
|
+
# @param min_periods [Integer]
|
3163
|
+
# The number of values in the window that should be non-null before computing
|
3164
|
+
# a result. If None, it will be set equal to window size.
|
3165
|
+
# @param center [Boolean]
|
3166
|
+
# Set the labels at the center of the window
|
3167
|
+
#
|
3168
|
+
# @return [Series]
|
3169
|
+
#
|
3170
|
+
# @example
|
3171
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, 4.0, 6.0, 8.0])
|
3172
|
+
# s.rolling_std(3)
|
3173
|
+
# # =>
|
3174
|
+
# # shape: (6,)
|
3175
|
+
# # Series: 'a' [f64]
|
3176
|
+
# # [
|
3177
|
+
# # null
|
3178
|
+
# # null
|
3179
|
+
# # 1.0
|
3180
|
+
# # 1.0
|
3181
|
+
# # 1.527525
|
3182
|
+
# # 2.0
|
3183
|
+
# # ]
|
3184
|
+
def rolling_std(
|
3185
|
+
window_size,
|
3186
|
+
weights: nil,
|
3187
|
+
min_periods: nil,
|
3188
|
+
center: false,
|
3189
|
+
ddof: 1
|
3190
|
+
)
|
3191
|
+
super
|
3192
|
+
end
|
3193
|
+
|
3194
|
+
# Compute a rolling variance.
|
3195
|
+
#
|
3196
|
+
# A window of length `window_size` will traverse the array. The values that fill
|
3197
|
+
# this window will (optionally) be multiplied with the weights given by the
|
3198
|
+
# `weight` vector. The resulting values will be aggregated to their sum.
|
3199
|
+
#
|
3200
|
+
# @param window_size [Integer]
|
3201
|
+
# The length of the window.
|
3202
|
+
# @param weights [Array]
|
3203
|
+
# An optional slice with the same length as the window that will be multiplied
|
3204
|
+
# elementwise with the values in the window.
|
3205
|
+
# @param min_periods [Integer]
|
3206
|
+
# The number of values in the window that should be non-null before computing
|
3207
|
+
# a result. If None, it will be set equal to window size.
|
3208
|
+
# @param center [Boolean]
|
3209
|
+
# Set the labels at the center of the window
|
3210
|
+
#
|
3211
|
+
# @return [Series]
|
3212
|
+
#
|
3213
|
+
# @example
|
3214
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, 4.0, 6.0, 8.0])
|
3215
|
+
# s.rolling_var(3)
|
3216
|
+
# # =>
|
3217
|
+
# # shape: (6,)
|
3218
|
+
# # Series: 'a' [f64]
|
3219
|
+
# # [
|
3220
|
+
# # null
|
3221
|
+
# # null
|
3222
|
+
# # 1.0
|
3223
|
+
# # 1.0
|
3224
|
+
# # 2.333333
|
3225
|
+
# # 4.0
|
3226
|
+
# # ]
|
3227
|
+
def rolling_var(
|
3228
|
+
window_size,
|
3229
|
+
weights: nil,
|
3230
|
+
min_periods: nil,
|
3231
|
+
center: false,
|
3232
|
+
ddof: 1
|
3233
|
+
)
|
3234
|
+
super
|
3235
|
+
end
|
3236
|
+
|
3237
|
+
# def rolling_apply
|
3238
|
+
# end
|
3239
|
+
|
3240
|
+
# Compute a rolling median.
|
3241
|
+
#
|
3242
|
+
# @param window_size [Integer]
|
3243
|
+
# The length of the window.
|
3244
|
+
# @param weights [Array]
|
3245
|
+
# An optional slice with the same length as the window that will be multiplied
|
3246
|
+
# elementwise with the values in the window.
|
3247
|
+
# @param min_periods [Integer]
|
3248
|
+
# The number of values in the window that should be non-null before computing
|
3249
|
+
# a result. If None, it will be set equal to window size.
|
3250
|
+
# @param center [Boolean]
|
3251
|
+
# Set the labels at the center of the window
|
3252
|
+
#
|
3253
|
+
# @return [Series]
|
3254
|
+
#
|
3255
|
+
# @example
|
3256
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, 4.0, 6.0, 8.0])
|
3257
|
+
# s.rolling_median(3)
|
3258
|
+
# # =>
|
3259
|
+
# # shape: (6,)
|
3260
|
+
# # Series: 'a' [f64]
|
3261
|
+
# # [
|
3262
|
+
# # null
|
3263
|
+
# # null
|
3264
|
+
# # 2.0
|
3265
|
+
# # 3.0
|
3266
|
+
# # 4.0
|
3267
|
+
# # 6.0
|
3268
|
+
# # ]
|
3269
|
+
def rolling_median(
|
3270
|
+
window_size,
|
3271
|
+
weights: nil,
|
3272
|
+
min_periods: nil,
|
3273
|
+
center: false
|
3274
|
+
)
|
3275
|
+
super
|
3276
|
+
end
|
3277
|
+
|
3278
|
+
# Compute a rolling quantile.
|
3279
|
+
#
|
3280
|
+
# @param quantile [Float]
|
3281
|
+
# Quantile between 0.0 and 1.0.
|
3282
|
+
# @param interpolation ["nearest", "higher", "lower", "midpoint", "linear"]
|
3283
|
+
# Interpolation method.
|
3284
|
+
# @param window_size [Integer]
|
3285
|
+
# The length of the window.
|
3286
|
+
# @param weights [Array]
|
3287
|
+
# An optional slice with the same length as the window that will be multiplied
|
3288
|
+
# elementwise with the values in the window.
|
3289
|
+
# @param min_periods [Integer]
|
3290
|
+
# The number of values in the window that should be non-null before computing
|
3291
|
+
# a result. If None, it will be set equal to window size.
|
3292
|
+
# @param center [Boolean]
|
3293
|
+
# Set the labels at the center of the window
|
3294
|
+
#
|
3295
|
+
# @return [Series]
|
3296
|
+
#
|
3297
|
+
# @example
|
3298
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, 4.0, 6.0, 8.0])
|
3299
|
+
# s.rolling_quantile(0.33, window_size: 3)
|
3300
|
+
# # =>
|
3301
|
+
# # shape: (6,)
|
3302
|
+
# # Series: 'a' [f64]
|
3303
|
+
# # [
|
3304
|
+
# # null
|
3305
|
+
# # null
|
3306
|
+
# # 1.0
|
3307
|
+
# # 2.0
|
3308
|
+
# # 3.0
|
3309
|
+
# # 4.0
|
3310
|
+
# # ]
|
3311
|
+
#
|
3312
|
+
# @example
|
3313
|
+
# s.rolling_quantile(0.33, interpolation: "linear", window_size: 3)
|
3314
|
+
# # =>
|
3315
|
+
# # shape: (6,)
|
3316
|
+
# # Series: 'a' [f64]
|
3317
|
+
# # [
|
3318
|
+
# # null
|
3319
|
+
# # null
|
3320
|
+
# # 1.66
|
3321
|
+
# # 2.66
|
3322
|
+
# # 3.66
|
3323
|
+
# # 5.32
|
3324
|
+
# # ]
|
3325
|
+
def rolling_quantile(
|
3326
|
+
quantile,
|
3327
|
+
interpolation: "nearest",
|
3328
|
+
window_size: 2,
|
3329
|
+
weights: nil,
|
3330
|
+
min_periods: nil,
|
3331
|
+
center: false
|
3332
|
+
)
|
3333
|
+
super
|
3334
|
+
end
|
3335
|
+
|
3336
|
+
# Compute a rolling skew.
|
3337
|
+
#
|
3338
|
+
# @param window_size [Integer]
|
3339
|
+
# Integer size of the rolling window.
|
3340
|
+
# @param bias [Boolean]
|
3341
|
+
# If false, the calculations are corrected for statistical bias.
|
3342
|
+
#
|
3343
|
+
# @return [Series]
|
3344
|
+
#
|
3345
|
+
# @example
|
3346
|
+
# s = Polars::Series.new("a", [1.0, 2.0, 3.0, 4.0, 6.0, 8.0])
|
3347
|
+
# s.rolling_skew(3)
|
3348
|
+
# # =>
|
3349
|
+
# # shape: (6,)
|
3350
|
+
# # Series: 'a' [f64]
|
3351
|
+
# # [
|
3352
|
+
# # null
|
3353
|
+
# # null
|
3354
|
+
# # 0.0
|
3355
|
+
# # 0.0
|
3356
|
+
# # 0.381802
|
3357
|
+
# # 0.0
|
3358
|
+
# # ]
|
3359
|
+
def rolling_skew(window_size, bias: true)
|
3360
|
+
super
|
3361
|
+
end
|
3362
|
+
|
3363
|
+
# Sample from this Series.
|
3364
|
+
#
|
3365
|
+
# @param n [Integer]
|
3366
|
+
# Number of items to return. Cannot be used with `frac`. Defaults to 1 if
|
3367
|
+
# `frac` is None.
|
3368
|
+
# @param frac [Float]
|
3369
|
+
# Fraction of items to return. Cannot be used with `n`.
|
3370
|
+
# @param with_replacement [Boolean]
|
3371
|
+
# Allow values to be sampled more than once.
|
3372
|
+
# @param shuffle [Boolean]
|
3373
|
+
# Shuffle the order of sampled data points.
|
3374
|
+
# @param seed [Integer]
|
3375
|
+
# Seed for the random number generator. If set to None (default), a random
|
3376
|
+
# seed is used.
|
3377
|
+
#
|
3378
|
+
# @return [Series]
|
3379
|
+
#
|
3380
|
+
# @example
|
3381
|
+
# s = Polars::Series.new("a", [1, 2, 3, 4, 5])
|
3382
|
+
# s.sample(n: 2, seed: 0)
|
3383
|
+
# # =>
|
3384
|
+
# # shape: (2,)
|
3385
|
+
# # Series: 'a' [i64]
|
3386
|
+
# # [
|
3387
|
+
# # 5
|
3388
|
+
# # 3
|
3389
|
+
# # ]
|
3390
|
+
def sample(
|
3391
|
+
n: nil,
|
3392
|
+
frac: nil,
|
3393
|
+
with_replacement: false,
|
3394
|
+
shuffle: false,
|
3395
|
+
seed: nil
|
3396
|
+
)
|
3397
|
+
if !n.nil? && !frac.nil?
|
3398
|
+
raise ArgumentError, "cannot specify both `n` and `frac`"
|
3399
|
+
end
|
3400
|
+
|
3401
|
+
if n.nil? && !frac.nil?
|
3402
|
+
return Utils.wrap_s(_s.sample_frac(frac, with_replacement, shuffle, seed))
|
3403
|
+
end
|
3404
|
+
|
3405
|
+
if n.nil?
|
3406
|
+
n = 1
|
3407
|
+
end
|
3408
|
+
Utils.wrap_s(_s.sample_n(n, with_replacement, shuffle, seed))
|
3409
|
+
end
|
3410
|
+
|
3411
|
+
# Get a boolean mask of the local maximum peaks.
|
3412
|
+
#
|
3413
|
+
# @return [Series]
|
3414
|
+
#
|
3415
|
+
# @example
|
3416
|
+
# s = Polars::Series.new("a", [1, 2, 3, 4, 5])
|
3417
|
+
# s.peak_max
|
3418
|
+
# # =>
|
3419
|
+
# # shape: (5,)
|
3420
|
+
# # Series: 'a' [bool]
|
3421
|
+
# # [
|
3422
|
+
# # false
|
3423
|
+
# # false
|
3424
|
+
# # false
|
3425
|
+
# # false
|
3426
|
+
# # true
|
3427
|
+
# # ]
|
3428
|
+
def peak_max
|
3429
|
+
super
|
3430
|
+
end
|
3431
|
+
|
3432
|
+
# Get a boolean mask of the local minimum peaks.
|
3433
|
+
#
|
3434
|
+
# @return [Series]
|
3435
|
+
#
|
3436
|
+
# @example
|
3437
|
+
# s = Polars::Series.new("a", [4, 1, 3, 2, 5])
|
3438
|
+
# s.peak_min
|
3439
|
+
# # =>
|
3440
|
+
# # shape: (5,)
|
3441
|
+
# # Series: 'a' [bool]
|
3442
|
+
# # [
|
3443
|
+
# # false
|
3444
|
+
# # true
|
3445
|
+
# # false
|
3446
|
+
# # true
|
3447
|
+
# # false
|
3448
|
+
# # ]
|
3449
|
+
def peak_min
|
3450
|
+
super
|
3451
|
+
end
|
3452
|
+
|
3453
|
+
# Count the number of unique values in this Series.
|
3454
|
+
#
|
3455
|
+
# @return [Integer]
|
3456
|
+
#
|
3457
|
+
# @example
|
3458
|
+
# s = Polars::Series.new("a", [1, 2, 2, 3])
|
3459
|
+
# s.n_unique
|
3460
|
+
# # => 3
|
3461
|
+
def n_unique
|
3462
|
+
_s.n_unique
|
3463
|
+
end
|
3464
|
+
|
3465
|
+
# Shrink Series memory usage.
|
3466
|
+
#
|
3467
|
+
# Shrinks the underlying array capacity to exactly fit the actual data.
|
3468
|
+
# (Note that this function does not change the Series data type).
|
3469
|
+
#
|
3470
|
+
# @return [Series]
|
3471
|
+
def shrink_to_fit(in_place: false)
|
3472
|
+
if in_place
|
3473
|
+
_s.shrink_to_fit
|
3474
|
+
self
|
3475
|
+
else
|
3476
|
+
series = clone
|
3477
|
+
series._s.shrink_to_fit
|
3478
|
+
series
|
3479
|
+
end
|
3480
|
+
end
|
3481
|
+
|
3482
|
+
# Hash the Series.
|
3483
|
+
#
|
3484
|
+
# The hash value is of type `:u64`.
|
3485
|
+
#
|
3486
|
+
# @param seed [Integer]
|
3487
|
+
# Random seed parameter. Defaults to 0.
|
3488
|
+
# @param seed_1 [Integer]
|
3489
|
+
# Random seed parameter. Defaults to `seed` if not set.
|
3490
|
+
# @param seed_2 [Integer]
|
3491
|
+
# Random seed parameter. Defaults to `seed` if not set.
|
3492
|
+
# @param seed_3 [Integer]
|
3493
|
+
# Random seed parameter. Defaults to `seed` if not set.
|
3494
|
+
#
|
3495
|
+
# @return [Series]
|
3496
|
+
#
|
3497
|
+
# @example
|
3498
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
3499
|
+
# s._hash(42)
|
3500
|
+
# # =>
|
3501
|
+
# # shape: (3,)
|
3502
|
+
# # Series: 'a' [u64]
|
3503
|
+
# # [
|
3504
|
+
# # 2374023516666777365
|
3505
|
+
# # 10386026231460783898
|
3506
|
+
# # 17796317186427479491
|
3507
|
+
# # ]
|
3508
|
+
def _hash(seed = 0, seed_1 = nil, seed_2 = nil, seed_3 = nil)
|
3509
|
+
super
|
3510
|
+
end
|
3511
|
+
|
3512
|
+
# Reinterpret the underlying bits as a signed/unsigned integer.
|
3513
|
+
#
|
3514
|
+
# This operation is only allowed for 64bit integers. For lower bits integers,
|
3515
|
+
# you can safely use that cast operation.
|
3516
|
+
#
|
3517
|
+
# @param signed [Boolean]
|
3518
|
+
# If true, reinterpret as `:i64`. Otherwise, reinterpret as `:u64`.
|
3519
|
+
#
|
3520
|
+
# @return [Series]
|
3521
|
+
def reinterpret(signed: true)
|
3522
|
+
super
|
3523
|
+
end
|
3524
|
+
|
3525
|
+
# Interpolate intermediate values. The interpolation method is linear.
|
3526
|
+
#
|
3527
|
+
# @return [Series]
|
3528
|
+
#
|
3529
|
+
# @example
|
3530
|
+
# s = Polars::Series.new("a", [1, 2, nil, nil, 5])
|
3531
|
+
# s.interpolate
|
3532
|
+
# # =>
|
3533
|
+
# # shape: (5,)
|
3534
|
+
# # Series: 'a' [f64]
|
3535
|
+
# # [
|
3536
|
+
# # 1.0
|
3537
|
+
# # 2.0
|
3538
|
+
# # 3.0
|
3539
|
+
# # 4.0
|
3540
|
+
# # 5.0
|
3541
|
+
# # ]
|
3542
|
+
def interpolate(method: "linear")
|
3543
|
+
super
|
3544
|
+
end
|
3545
|
+
|
3546
|
+
# Compute absolute values.
|
3547
|
+
#
|
3548
|
+
# @return [Series]
|
3549
|
+
def abs
|
3550
|
+
super
|
3551
|
+
end
|
3552
|
+
|
3553
|
+
# Assign ranks to data, dealing with ties appropriately.
|
3554
|
+
#
|
3555
|
+
# @param method ["average", "min", "max", "dense", "ordinal", "random"]
|
3556
|
+
# The method used to assign ranks to tied elements.
|
3557
|
+
# The following methods are available (default is 'average'):
|
3558
|
+
#
|
3559
|
+
# - 'average' : The average of the ranks that would have been assigned to
|
3560
|
+
# all the tied values is assigned to each value.
|
3561
|
+
# - 'min' : The minimum of the ranks that would have been assigned to all
|
3562
|
+
# the tied values is assigned to each value. (This is also referred to
|
3563
|
+
# as "competition" ranking.)
|
3564
|
+
# - 'max' : The maximum of the ranks that would have been assigned to all
|
3565
|
+
# the tied values is assigned to each value.
|
3566
|
+
# - 'dense' : Like 'min', but the rank of the next highest element is
|
3567
|
+
# assigned the rank immediately after those assigned to the tied
|
3568
|
+
# elements.
|
3569
|
+
# - 'ordinal' : All values are given a distinct rank, corresponding to
|
3570
|
+
# the order that the values occur in the Series.
|
3571
|
+
# - 'random' : Like 'ordinal', but the rank for ties is not dependent
|
3572
|
+
# on the order that the values occur in the Series.
|
3573
|
+
# @param reverse [Boolean]
|
3574
|
+
# Reverse the operation.
|
3575
|
+
# @param seed [Integer]
|
3576
|
+
# If `method: "random"`, use this as seed.
|
3577
|
+
#
|
3578
|
+
# @return [Series]
|
3579
|
+
#
|
3580
|
+
# @example The 'average' method:
|
3581
|
+
# s = Polars::Series.new("a", [3, 6, 1, 1, 6])
|
3582
|
+
# s.rank
|
3583
|
+
# # =>
|
3584
|
+
# # shape: (5,)
|
3585
|
+
# # Series: 'a' [f64]
|
3586
|
+
# # [
|
3587
|
+
# # 3.0
|
3588
|
+
# # 4.5
|
3589
|
+
# # 1.5
|
3590
|
+
# # 1.5
|
3591
|
+
# # 4.5
|
3592
|
+
# # ]
|
3593
|
+
#
|
3594
|
+
# @example The 'ordinal' method:
|
3595
|
+
# s = Polars::Series.new("a", [3, 6, 1, 1, 6])
|
3596
|
+
# s.rank(method: "ordinal")
|
3597
|
+
# # =>
|
3598
|
+
# # shape: (5,)
|
3599
|
+
# # Series: 'a' [u32]
|
3600
|
+
# # [
|
3601
|
+
# # 3
|
3602
|
+
# # 4
|
3603
|
+
# # 1
|
3604
|
+
# # 2
|
3605
|
+
# # 5
|
3606
|
+
# # ]
|
3607
|
+
def rank(method: "average", reverse: false, seed: nil)
|
3608
|
+
super
|
3609
|
+
end
|
3610
|
+
|
3611
|
+
# Calculate the n-th discrete difference.
|
3612
|
+
#
|
3613
|
+
# @param n [Integer]
|
3614
|
+
# Number of slots to shift.
|
3615
|
+
# @param null_behavior ["ignore", "drop"]
|
3616
|
+
# How to handle null values.
|
3617
|
+
#
|
3618
|
+
# @return [Series]
|
3619
|
+
def diff(n: 1, null_behavior: "ignore")
|
3620
|
+
super
|
3621
|
+
end
|
3622
|
+
|
3623
|
+
# Computes percentage change between values.
|
3624
|
+
#
|
3625
|
+
# Percentage change (as fraction) between current element and most-recent
|
3626
|
+
# non-null element at least `n` period(s) before the current element.
|
3627
|
+
#
|
3628
|
+
# Computes the change from the previous row by default.
|
3629
|
+
#
|
3630
|
+
# @param n [Integer]
|
3631
|
+
# periods to shift for forming percent change.
|
3632
|
+
#
|
3633
|
+
# @return [Series]
|
3634
|
+
#
|
3635
|
+
# @example
|
3636
|
+
# Polars::Series.new(0..9).pct_change
|
3637
|
+
# # =>
|
3638
|
+
# # shape: (10,)
|
3639
|
+
# # Series: '' [f64]
|
3640
|
+
# # [
|
3641
|
+
# # null
|
3642
|
+
# # inf
|
3643
|
+
# # 1.0
|
3644
|
+
# # 0.5
|
3645
|
+
# # 0.333333
|
3646
|
+
# # 0.25
|
3647
|
+
# # 0.2
|
3648
|
+
# # 0.166667
|
3649
|
+
# # 0.142857
|
3650
|
+
# # 0.125
|
3651
|
+
# # ]
|
3652
|
+
#
|
3653
|
+
# @example
|
3654
|
+
# Polars::Series.new([1, 2, 4, 8, 16, 32, 64, 128, 256, 512]).pct_change(n: 2)
|
3655
|
+
# # =>
|
3656
|
+
# # shape: (10,)
|
3657
|
+
# # Series: '' [f64]
|
3658
|
+
# # [
|
3659
|
+
# # null
|
3660
|
+
# # null
|
3661
|
+
# # 3.0
|
3662
|
+
# # 3.0
|
3663
|
+
# # 3.0
|
3664
|
+
# # 3.0
|
3665
|
+
# # 3.0
|
3666
|
+
# # 3.0
|
3667
|
+
# # 3.0
|
3668
|
+
# # 3.0
|
3669
|
+
# # ]
|
3670
|
+
def pct_change(n: 1)
|
3671
|
+
super
|
3672
|
+
end
|
3673
|
+
|
3674
|
+
# Compute the sample skewness of a data set.
|
3675
|
+
#
|
3676
|
+
# For normally distributed data, the skewness should be about zero. For
|
3677
|
+
# unimodal continuous distributions, a skewness value greater than zero means
|
3678
|
+
# that there is more weight in the right tail of the distribution. The
|
3679
|
+
# function `skewtest` can be used to determine if the skewness value
|
3680
|
+
# is close enough to zero, statistically speaking.
|
3681
|
+
#
|
3682
|
+
# @param bias [Boolean]
|
3683
|
+
# If `false`, the calculations are corrected for statistical bias.
|
3684
|
+
#
|
3685
|
+
# @return [Float, nil]
|
3686
|
+
def skew(bias: true)
|
3687
|
+
_s.skew(bias)
|
3688
|
+
end
|
3689
|
+
|
3690
|
+
# Compute the kurtosis (Fisher or Pearson) of a dataset.
|
3691
|
+
#
|
3692
|
+
# Kurtosis is the fourth central moment divided by the square of the
|
3693
|
+
# variance. If Fisher's definition is used, then 3.0 is subtracted from
|
3694
|
+
# the result to give 0.0 for a normal distribution.
|
3695
|
+
# If bias is false, then the kurtosis is calculated using k statistics to
|
3696
|
+
# eliminate bias coming from biased moment estimators
|
3697
|
+
#
|
3698
|
+
# @param fisher [Boolean]
|
3699
|
+
# If `true`, Fisher's definition is used (normal ==> 0.0). If `false`,
|
3700
|
+
# Pearson's definition is used (normal ==> 3.0).
|
3701
|
+
# @param bias [Boolean]
|
3702
|
+
# If `false`, the calculations are corrected for statistical bias.
|
3703
|
+
#
|
3704
|
+
# @return [Float, nil]
|
3705
|
+
def kurtosis(fisher: true, bias: true)
|
3706
|
+
_s.kurtosis(fisher, bias)
|
3707
|
+
end
|
3708
|
+
|
3709
|
+
# Clip (limit) the values in an array to a `min` and `max` boundary.
|
3710
|
+
#
|
3711
|
+
# Only works for numerical types.
|
3712
|
+
#
|
3713
|
+
# If you want to clip other dtypes, consider writing a "when, then, otherwise"
|
3714
|
+
# expression. See {#when} for more information.
|
3715
|
+
#
|
3716
|
+
# @param min_val [Numeric]
|
3717
|
+
# Minimum value.
|
3718
|
+
# @param max_val [Numeric]
|
3719
|
+
# Maximum value.
|
3720
|
+
#
|
3721
|
+
# @return [Series]
|
3722
|
+
#
|
3723
|
+
# @example
|
3724
|
+
# s = Polars::Series.new("foo", [-50, 5, nil, 50])
|
3725
|
+
# s.clip(1, 10)
|
3726
|
+
# # =>
|
3727
|
+
# # shape: (4,)
|
3728
|
+
# # Series: 'foo' [i64]
|
3729
|
+
# # [
|
3730
|
+
# # 1
|
3731
|
+
# # 5
|
3732
|
+
# # null
|
3733
|
+
# # 10
|
3734
|
+
# # ]
|
3735
|
+
def clip(min_val = nil, max_val = nil)
|
3736
|
+
super
|
3737
|
+
end
|
3738
|
+
|
3739
|
+
# Clip (limit) the values in an array to a `min` boundary.
|
3740
|
+
#
|
3741
|
+
# Only works for numerical types.
|
3742
|
+
#
|
3743
|
+
# If you want to clip other dtypes, consider writing a "when, then, otherwise"
|
3744
|
+
# expression. See {#when} for more information.
|
3745
|
+
#
|
3746
|
+
# @param min_val [Numeric]
|
3747
|
+
# Minimum value.
|
3748
|
+
#
|
3749
|
+
# @return [Series]
|
3750
|
+
def clip_min(min_val)
|
3751
|
+
super
|
3752
|
+
end
|
3753
|
+
|
3754
|
+
# Clip (limit) the values in an array to a `max` boundary.
|
3755
|
+
#
|
3756
|
+
# Only works for numerical types.
|
3757
|
+
#
|
3758
|
+
# If you want to clip other dtypes, consider writing a "when, then, otherwise"
|
3759
|
+
# expression. See {#when} for more information.
|
3760
|
+
#
|
3761
|
+
# @param max_val [Numeric]
|
3762
|
+
# Maximum value.
|
3763
|
+
#
|
3764
|
+
# @return [Series]
|
3765
|
+
def clip_max(max_val)
|
3766
|
+
super
|
3767
|
+
end
|
3768
|
+
|
3769
|
+
# Replace values by different values.
|
3770
|
+
#
|
3771
|
+
# @param old [Object]
|
3772
|
+
# Value or sequence of values to replace.
|
3773
|
+
# Also accepts a mapping of values to their replacement.
|
3774
|
+
# @param new [Object]
|
3775
|
+
# Value or sequence of values to replace by.
|
3776
|
+
# Length must match the length of `old` or have length 1.
|
3777
|
+
# @param default [Object]
|
3778
|
+
# Set values that were not replaced to this value.
|
3779
|
+
# Defaults to keeping the original value.
|
3780
|
+
# Accepts expression input. Non-expression inputs are parsed as literals.
|
3781
|
+
# @param return_dtype [Object]
|
3782
|
+
# The data type of the resulting Series. If set to `nil` (default),
|
3783
|
+
# the data type is determined automatically based on the other inputs.
|
3784
|
+
#
|
3785
|
+
# @return [Series]
|
3786
|
+
#
|
3787
|
+
# @example Replace a single value by another value. Values that were not replaced remain unchanged.
|
3788
|
+
# s = Polars::Series.new([1, 2, 2, 3])
|
3789
|
+
# s.replace(2, 100)
|
3790
|
+
# # =>
|
3791
|
+
# # shape: (4,)
|
3792
|
+
# # Series: '' [i64]
|
3793
|
+
# # [
|
3794
|
+
# # 1
|
3795
|
+
# # 100
|
3796
|
+
# # 100
|
3797
|
+
# # 3
|
3798
|
+
# # ]
|
3799
|
+
#
|
3800
|
+
# @example Replace multiple values by passing sequences to the `old` and `new` parameters.
|
3801
|
+
# s.replace([2, 3], [100, 200])
|
3802
|
+
# # =>
|
3803
|
+
# # shape: (4,)
|
3804
|
+
# # Series: '' [i64]
|
3805
|
+
# # [
|
3806
|
+
# # 1
|
3807
|
+
# # 100
|
3808
|
+
# # 100
|
3809
|
+
# # 200
|
3810
|
+
# # ]
|
3811
|
+
#
|
3812
|
+
# @example Passing a mapping with replacements is also supported as syntactic sugar.
|
3813
|
+
# mapping = {2 => 100, 3 => 200}
|
3814
|
+
# s.replace(mapping)
|
3815
|
+
# # =>
|
3816
|
+
# # shape: (4,)
|
3817
|
+
# # Series: '' [i64]
|
3818
|
+
# # [
|
3819
|
+
# # 1
|
3820
|
+
# # 100
|
3821
|
+
# # 100
|
3822
|
+
# # 200
|
3823
|
+
# # ]
|
3824
|
+
#
|
3825
|
+
# @example The original data type is preserved when replacing by values of a different data type.
|
3826
|
+
# s = Polars::Series.new(["x", "y", "z"])
|
3827
|
+
# mapping = {"x" => 1, "y" => 2, "z" => 3}
|
3828
|
+
# s.replace(mapping)
|
3829
|
+
# # =>
|
3830
|
+
# # shape: (3,)
|
3831
|
+
# # Series: '' [str]
|
3832
|
+
# # [
|
3833
|
+
# # "1"
|
3834
|
+
# # "2"
|
3835
|
+
# # "3"
|
3836
|
+
# # ]
|
3837
|
+
def replace(old, new = Expr::NO_DEFAULT, default: Expr::NO_DEFAULT, return_dtype: nil)
|
3838
|
+
super
|
3839
|
+
end
|
3840
|
+
|
3841
|
+
# Reshape this Series to a flat Series or a Series of Lists.
|
3842
|
+
#
|
3843
|
+
# @param dims [Array]
|
3844
|
+
# Tuple of the dimension sizes. If a -1 is used in any of the dimensions, that
|
3845
|
+
# dimension is inferred.
|
3846
|
+
#
|
3847
|
+
# @return [Series]
|
3848
|
+
def reshape(dims)
|
3849
|
+
super
|
3850
|
+
end
|
3851
|
+
|
3852
|
+
# Shuffle the contents of this Series.
|
3853
|
+
#
|
3854
|
+
# @param seed [Integer, nil]
|
3855
|
+
# Seed for the random number generator.
|
3856
|
+
#
|
3857
|
+
# @return [Series]
|
3858
|
+
#
|
3859
|
+
# @example
|
3860
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
3861
|
+
# s.shuffle(seed: 1)
|
3862
|
+
# # =>
|
3863
|
+
# # shape: (3,)
|
3864
|
+
# # Series: 'a' [i64]
|
3865
|
+
# # [
|
3866
|
+
# # 2
|
3867
|
+
# # 1
|
3868
|
+
# # 3
|
3869
|
+
# # ]
|
3870
|
+
def shuffle(seed: nil)
|
3871
|
+
super
|
3872
|
+
end
|
3873
|
+
|
3874
|
+
# Exponentially-weighted moving average.
|
3875
|
+
#
|
3876
|
+
# @return [Series]
|
3877
|
+
def ewm_mean(
|
3878
|
+
com: nil,
|
3879
|
+
span: nil,
|
3880
|
+
half_life: nil,
|
3881
|
+
alpha: nil,
|
3882
|
+
adjust: true,
|
3883
|
+
min_periods: 1,
|
3884
|
+
ignore_nulls: true
|
3885
|
+
)
|
3886
|
+
super
|
3887
|
+
end
|
3888
|
+
|
3889
|
+
# Exponentially-weighted moving standard deviation.
|
3890
|
+
#
|
3891
|
+
# @return [Series]
|
3892
|
+
def ewm_std(
|
3893
|
+
com: nil,
|
3894
|
+
span: nil,
|
3895
|
+
half_life: nil,
|
3896
|
+
alpha: nil,
|
3897
|
+
adjust: true,
|
3898
|
+
bias: false,
|
3899
|
+
min_periods: 1,
|
3900
|
+
ignore_nulls: true
|
3901
|
+
)
|
3902
|
+
super
|
3903
|
+
end
|
3904
|
+
|
3905
|
+
# Exponentially-weighted moving variance.
|
3906
|
+
#
|
3907
|
+
# @return [Series]
|
3908
|
+
def ewm_var(
|
3909
|
+
com: nil,
|
3910
|
+
span: nil,
|
3911
|
+
half_life: nil,
|
3912
|
+
alpha: nil,
|
3913
|
+
adjust: true,
|
3914
|
+
bias: false,
|
3915
|
+
min_periods: 1,
|
3916
|
+
ignore_nulls: true
|
3917
|
+
)
|
3918
|
+
super
|
3919
|
+
end
|
3920
|
+
|
3921
|
+
# Extend the Series with given number of values.
|
3922
|
+
#
|
3923
|
+
# @param value [Object]
|
3924
|
+
# The value to extend the Series with. This value may be `nil` to fill with
|
3925
|
+
# nulls.
|
3926
|
+
# @param n [Integer]
|
3927
|
+
# The number of values to extend.
|
3928
|
+
#
|
3929
|
+
# @return [Series]
|
3930
|
+
#
|
3931
|
+
# @example
|
3932
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
3933
|
+
# s.extend_constant(99, 2)
|
3934
|
+
# # =>
|
3935
|
+
# # shape: (5,)
|
3936
|
+
# # Series: 'a' [i64]
|
3937
|
+
# # [
|
3938
|
+
# # 1
|
3939
|
+
# # 2
|
3940
|
+
# # 3
|
3941
|
+
# # 99
|
3942
|
+
# # 99
|
3943
|
+
# # ]
|
3944
|
+
def extend_constant(value, n)
|
3945
|
+
Utils.wrap_s(_s.extend_constant(value, n))
|
3946
|
+
end
|
3947
|
+
|
3948
|
+
# Flags the Series as sorted.
|
3949
|
+
#
|
3950
|
+
# Enables downstream code to user fast paths for sorted arrays.
|
3951
|
+
#
|
3952
|
+
# @param reverse [Boolean]
|
3953
|
+
# If the Series order is reversed, e.g. descending.
|
3954
|
+
#
|
3955
|
+
# @return [Series]
|
3956
|
+
#
|
3957
|
+
# @note
|
3958
|
+
# This can lead to incorrect results if this Series is not sorted!!
|
3959
|
+
# Use with care!
|
3960
|
+
#
|
3961
|
+
# @example
|
3962
|
+
# s = Polars::Series.new("a", [1, 2, 3])
|
3963
|
+
# s.set_sorted.max
|
3964
|
+
# # => 3
|
3965
|
+
def set_sorted(reverse: false)
|
3966
|
+
Utils.wrap_s(_s.set_sorted(reverse))
|
3967
|
+
end
|
3968
|
+
|
3969
|
+
# Create a new Series filled with values from the given index.
|
3970
|
+
#
|
3971
|
+
# @return [Series]
|
3972
|
+
def new_from_index(index, length)
|
3973
|
+
Utils.wrap_s(_s.new_from_index(index, length))
|
3974
|
+
end
|
3975
|
+
|
3976
|
+
# Shrink numeric columns to the minimal required datatype.
|
3977
|
+
#
|
3978
|
+
# Shrink to the dtype needed to fit the extrema of this Series.
|
3979
|
+
# This can be used to reduce memory pressure.
|
3980
|
+
#
|
3981
|
+
# @return [Series]
|
3982
|
+
def shrink_dtype
|
3983
|
+
super
|
3984
|
+
end
|
3985
|
+
|
3986
|
+
# Create an object namespace of all list related methods.
|
3987
|
+
#
|
3988
|
+
# @return [ListNameSpace]
|
3989
|
+
def list
|
3990
|
+
ListNameSpace.new(self)
|
3991
|
+
end
|
3992
|
+
|
3993
|
+
# Create an object namespace of all array related methods.
|
3994
|
+
#
|
3995
|
+
# @return [ArrayNameSpace]
|
3996
|
+
def arr
|
3997
|
+
ArrayNameSpace.new(self)
|
3998
|
+
end
|
3999
|
+
|
4000
|
+
# Create an object namespace of all binary related methods.
|
4001
|
+
#
|
4002
|
+
# @return [BinaryNameSpace]
|
4003
|
+
def bin
|
4004
|
+
BinaryNameSpace.new(self)
|
4005
|
+
end
|
4006
|
+
|
4007
|
+
# Create an object namespace of all categorical related methods.
|
4008
|
+
#
|
4009
|
+
# @return [CatNameSpace]
|
4010
|
+
def cat
|
4011
|
+
CatNameSpace.new(self)
|
4012
|
+
end
|
4013
|
+
|
4014
|
+
# Create an object namespace of all datetime related methods.
|
4015
|
+
#
|
4016
|
+
# @return [DateTimeNameSpace]
|
4017
|
+
def dt
|
4018
|
+
DateTimeNameSpace.new(self)
|
4019
|
+
end
|
4020
|
+
|
4021
|
+
# Create an object namespace of all string related methods.
|
4022
|
+
#
|
4023
|
+
# @return [StringNameSpace]
|
4024
|
+
def str
|
4025
|
+
StringNameSpace.new(self)
|
4026
|
+
end
|
4027
|
+
|
4028
|
+
# Create an object namespace of all struct related methods.
|
4029
|
+
#
|
4030
|
+
# @return [StructNameSpace]
|
4031
|
+
def struct
|
4032
|
+
StructNameSpace.new(self)
|
4033
|
+
end
|
4034
|
+
|
4035
|
+
private
|
4036
|
+
|
4037
|
+
def initialize_copy(other)
|
4038
|
+
super
|
4039
|
+
self._s = _s._clone
|
4040
|
+
end
|
4041
|
+
|
4042
|
+
def coerce(other)
|
4043
|
+
if other.is_a?(Numeric)
|
4044
|
+
# TODO improve
|
4045
|
+
series = to_frame.select(Polars.lit(other)).to_series
|
4046
|
+
[series, self]
|
4047
|
+
else
|
4048
|
+
raise TypeError, "#{self.class} can't be coerced into #{other.class}"
|
4049
|
+
end
|
4050
|
+
end
|
4051
|
+
|
4052
|
+
def _pos_idxs(idxs)
|
4053
|
+
idx_type = Plr.get_index_type
|
4054
|
+
|
4055
|
+
if idxs.is_a?(Series)
|
4056
|
+
if idxs.dtype == idx_type
|
4057
|
+
return idxs
|
4058
|
+
end
|
4059
|
+
if [UInt8, UInt16, idx_type == UInt32 ? UInt64 : UInt32, Int8, Int16, Int32, Int64].include?(idxs.dtype)
|
4060
|
+
if idx_type == UInt32
|
4061
|
+
if [Int64, UInt64].include?(idxs.dtype)
|
4062
|
+
if idxs.max >= 2**32
|
4063
|
+
raise ArgumentError, "Index positions should be smaller than 2^32."
|
4064
|
+
end
|
4065
|
+
end
|
4066
|
+
if idxs.dtype == Int64
|
4067
|
+
if idxs.min < -(2**32)
|
4068
|
+
raise ArgumentError, "Index positions should be bigger than -2^32 + 1."
|
4069
|
+
end
|
4070
|
+
end
|
4071
|
+
end
|
4072
|
+
if [Int8, Int16, Int32, Int64].include?(idxs.dtype)
|
4073
|
+
if idxs.min < 0
|
4074
|
+
if idx_type == UInt32
|
4075
|
+
if [Int8, Int16].include?(idxs.dtype)
|
4076
|
+
idxs = idxs.cast(Int32)
|
4077
|
+
end
|
4078
|
+
else
|
4079
|
+
if [Int8, Int16, Int32].include?(idxs.dtype)
|
4080
|
+
idxs = idxs.cast(Int64)
|
4081
|
+
end
|
4082
|
+
end
|
4083
|
+
|
4084
|
+
# Update negative indexes to absolute indexes.
|
4085
|
+
return (
|
4086
|
+
idxs.to_frame
|
4087
|
+
.select(
|
4088
|
+
Polars.when(Polars.col(idxs.name) < 0)
|
4089
|
+
.then(len + Polars.col(idxs.name))
|
4090
|
+
.otherwise(Polars.col(idxs.name))
|
4091
|
+
.cast(idx_type)
|
4092
|
+
)
|
4093
|
+
.to_series(0)
|
4094
|
+
)
|
4095
|
+
end
|
4096
|
+
end
|
4097
|
+
|
4098
|
+
return idxs.cast(idx_type)
|
4099
|
+
end
|
4100
|
+
end
|
4101
|
+
|
4102
|
+
raise ArgumentError, "Unsupported idxs datatype."
|
4103
|
+
end
|
4104
|
+
|
4105
|
+
def _comp(other, op)
|
4106
|
+
if dtype == Boolean && Utils.bool?(other) && [:eq, :neq].include?(op)
|
4107
|
+
if (other == true && op == :eq) || (other == false && op == :neq)
|
4108
|
+
return clone
|
4109
|
+
elsif (other == false && op == :eq) || (other == true && op == :neq)
|
4110
|
+
return !self
|
4111
|
+
end
|
4112
|
+
end
|
4113
|
+
|
4114
|
+
if other.is_a?(::Time) && dtype.is_a?(Datetime)
|
4115
|
+
ts = Utils.datetime_to_int(other, time_unit)
|
4116
|
+
f = ffi_func("#{op}_<>", Int64, _s)
|
4117
|
+
fail if f.nil?
|
4118
|
+
return Utils.wrap_s(f.call(ts))
|
4119
|
+
elsif other.is_a?(::Date) && dtype == Date
|
4120
|
+
d = Utils.date_to_int(other)
|
4121
|
+
f = ffi_func("#{op}_<>", Int32, _s)
|
4122
|
+
fail if f.nil?
|
4123
|
+
return Utils.wrap_s(f.call(d))
|
4124
|
+
end
|
4125
|
+
|
4126
|
+
if other.is_a?(Series)
|
4127
|
+
return Utils.wrap_s(_s.send(op, other._s))
|
4128
|
+
end
|
4129
|
+
|
4130
|
+
f = ffi_func("#{op}_<>", dtype, _s)
|
4131
|
+
if f.nil?
|
4132
|
+
raise NotImplementedError
|
4133
|
+
end
|
4134
|
+
Utils.wrap_s(f.call(other))
|
4135
|
+
end
|
4136
|
+
|
4137
|
+
def ffi_func(name, dtype, _s)
|
4138
|
+
_s.method(name.sub("<>", DTYPE_TO_FFINAME.fetch(dtype.class))) if DTYPE_TO_FFINAME.key?(dtype.class)
|
4139
|
+
end
|
4140
|
+
|
4141
|
+
def _arithmetic(other, op)
|
4142
|
+
if other.is_a?(Expr)
|
4143
|
+
other = to_frame.select(other).to_series
|
4144
|
+
end
|
4145
|
+
if other.is_a?(Series)
|
4146
|
+
return Utils.wrap_s(_s.send(op, other._s))
|
4147
|
+
end
|
4148
|
+
|
4149
|
+
if (other.is_a?(Float) || other.is_a?(::Date) || other.is_a?(::DateTime) || other.is_a?(::Time) || other.is_a?(::String)) && !is_float
|
4150
|
+
_s2 = sequence_to_rbseries(name, [other])
|
4151
|
+
return Utils.wrap_s(_s.send(op, _s2))
|
4152
|
+
end
|
4153
|
+
|
4154
|
+
f = ffi_func("#{op}_<>", dtype, _s)
|
4155
|
+
if f.nil?
|
4156
|
+
raise ArgumentError, "cannot do arithmetic with series of dtype: #{dtype} and argument of type: #{other.class.name}"
|
4157
|
+
end
|
4158
|
+
Utils.wrap_s(f.call(other))
|
4159
|
+
end
|
4160
|
+
|
4161
|
+
DTYPE_TO_FFINAME = {
|
4162
|
+
Int8 => "i8",
|
4163
|
+
Int16 => "i16",
|
4164
|
+
Int32 => "i32",
|
4165
|
+
Int64 => "i64",
|
4166
|
+
UInt8 => "u8",
|
4167
|
+
UInt16 => "u16",
|
4168
|
+
UInt32 => "u32",
|
4169
|
+
UInt64 => "u64",
|
4170
|
+
Float32 => "f32",
|
4171
|
+
Float64 => "f64",
|
4172
|
+
Boolean => "bool",
|
4173
|
+
Utf8 => "str",
|
4174
|
+
List => "list",
|
4175
|
+
Date => "date",
|
4176
|
+
Datetime => "datetime",
|
4177
|
+
Duration => "duration",
|
4178
|
+
Time => "time",
|
4179
|
+
Object => "object",
|
4180
|
+
Categorical => "categorical",
|
4181
|
+
Struct => "struct",
|
4182
|
+
Binary => "binary"
|
4183
|
+
}
|
4184
|
+
|
4185
|
+
def series_to_rbseries(name, values)
|
4186
|
+
# should not be in-place?
|
4187
|
+
values.rename(name, in_place: true)
|
4188
|
+
values._s
|
4189
|
+
end
|
4190
|
+
|
4191
|
+
def numo_to_rbseries(name, values, strict: true, nan_to_null: false)
|
4192
|
+
# not needed yet
|
4193
|
+
# if !values.contiguous?
|
4194
|
+
# end
|
4195
|
+
|
4196
|
+
if values.shape.length == 1
|
4197
|
+
values, dtype = numo_values_and_dtype(values)
|
4198
|
+
strict = nan_to_null if [Numo::SFloat, Numo::DFloat].include?(dtype)
|
4199
|
+
if dtype == Numo::RObject
|
4200
|
+
sequence_to_rbseries(name, values.to_a, strict: strict)
|
4201
|
+
else
|
4202
|
+
constructor = numo_type_to_constructor(dtype)
|
4203
|
+
# TODO improve performance
|
4204
|
+
constructor.call(name, values.to_a, strict)
|
4205
|
+
end
|
4206
|
+
elsif values.shape.sum == 0
|
4207
|
+
raise Todo
|
4208
|
+
else
|
4209
|
+
original_shape = values.shape
|
4210
|
+
values = values.reshape(original_shape.inject(&:*))
|
4211
|
+
rb_s = numo_to_rbseries(
|
4212
|
+
name,
|
4213
|
+
values,
|
4214
|
+
strict: strict,
|
4215
|
+
nan_to_null: nan_to_null
|
4216
|
+
)
|
4217
|
+
Utils.wrap_s(rb_s).reshape(original_shape)._s
|
4218
|
+
end
|
4219
|
+
end
|
4220
|
+
|
4221
|
+
def numo_values_and_dtype(values)
|
4222
|
+
[values, values.class]
|
4223
|
+
end
|
4224
|
+
|
4225
|
+
def numo_type_to_constructor(dtype)
|
4226
|
+
{
|
4227
|
+
Numo::Float32 => RbSeries.method(:new_opt_f32),
|
4228
|
+
Numo::Float64 => RbSeries.method(:new_opt_f64),
|
4229
|
+
Numo::Int8 => RbSeries.method(:new_opt_i8),
|
4230
|
+
Numo::Int16 => RbSeries.method(:new_opt_i16),
|
4231
|
+
Numo::Int32 => RbSeries.method(:new_opt_i32),
|
4232
|
+
Numo::Int64 => RbSeries.method(:new_opt_i64),
|
4233
|
+
Numo::UInt8 => RbSeries.method(:new_opt_u8),
|
4234
|
+
Numo::UInt16 => RbSeries.method(:new_opt_u16),
|
4235
|
+
Numo::UInt32 => RbSeries.method(:new_opt_u32),
|
4236
|
+
Numo::UInt64 => RbSeries.method(:new_opt_u64)
|
4237
|
+
}.fetch(dtype)
|
4238
|
+
rescue KeyError
|
4239
|
+
RbSeries.method(:new_object)
|
4240
|
+
end
|
4241
|
+
|
4242
|
+
def sequence_to_rbseries(name, values, dtype: nil, strict: true, dtype_if_empty: nil)
|
4243
|
+
ruby_dtype = nil
|
4244
|
+
|
4245
|
+
if (values.nil? || values.empty?) && dtype.nil?
|
4246
|
+
dtype = dtype_if_empty || Float32
|
4247
|
+
elsif dtype == List
|
4248
|
+
ruby_dtype = ::Array
|
4249
|
+
end
|
4250
|
+
|
4251
|
+
rb_temporal_types = [::Date, ::DateTime, ::Time]
|
4252
|
+
rb_temporal_types << ActiveSupport::TimeWithZone if defined?(ActiveSupport::TimeWithZone)
|
4253
|
+
|
4254
|
+
value = _get_first_non_none(values)
|
4255
|
+
if !value.nil?
|
4256
|
+
if value.is_a?(Hash)
|
4257
|
+
return DataFrame.new(values).to_struct(name)._s
|
4258
|
+
end
|
4259
|
+
end
|
4260
|
+
|
4261
|
+
if !dtype.nil? && ![List, Struct, Unknown].include?(dtype) && Utils.is_polars_dtype(dtype) && ruby_dtype.nil?
|
4262
|
+
if dtype == Array && !dtype.is_a?(Array) && value.is_a?(::Array)
|
4263
|
+
dtype = Array.new(nil, value.size)
|
4264
|
+
end
|
4265
|
+
|
4266
|
+
constructor = polars_type_to_constructor(dtype)
|
4267
|
+
# TODO remove
|
4268
|
+
strict = false if dtype == Decimal
|
4269
|
+
rbseries = constructor.call(name, values, strict)
|
4270
|
+
|
4271
|
+
base_type = dtype.is_a?(DataType) ? dtype.class : dtype
|
4272
|
+
if [Date, Datetime, Duration, Time, Categorical, Boolean, Enum].include?(base_type)
|
4273
|
+
if rbseries.dtype != dtype
|
4274
|
+
rbseries = rbseries.cast(dtype, true)
|
4275
|
+
end
|
4276
|
+
end
|
4277
|
+
rbseries
|
4278
|
+
elsif dtype == Struct
|
4279
|
+
struct_schema = dtype.is_a?(Struct) ? dtype.to_schema : nil
|
4280
|
+
empty = {}
|
4281
|
+
DataFrame.sequence_to_rbdf(
|
4282
|
+
values.map { |v| v.nil? ? empty : v },
|
4283
|
+
schema: struct_schema,
|
4284
|
+
orient: "row",
|
4285
|
+
).to_struct(name)
|
4286
|
+
else
|
4287
|
+
if ruby_dtype.nil?
|
4288
|
+
if value.nil?
|
4289
|
+
# generic default dtype
|
4290
|
+
ruby_dtype = Float
|
4291
|
+
else
|
4292
|
+
ruby_dtype = value.class
|
4293
|
+
end
|
4294
|
+
end
|
4295
|
+
|
4296
|
+
# temporal branch
|
4297
|
+
if rb_temporal_types.include?(ruby_dtype)
|
4298
|
+
if dtype.nil?
|
4299
|
+
dtype = Utils.rb_type_to_dtype(ruby_dtype)
|
4300
|
+
elsif rb_temporal_types.include?(dtype)
|
4301
|
+
dtype = Utils.rb_type_to_dtype(dtype)
|
4302
|
+
end
|
4303
|
+
# TODO
|
4304
|
+
time_unit = nil
|
4305
|
+
|
4306
|
+
rb_series = RbSeries.new_from_any_values(name, values, strict)
|
4307
|
+
if time_unit.nil?
|
4308
|
+
s = Utils.wrap_s(rb_series)
|
4309
|
+
else
|
4310
|
+
s = Utils.wrap_s(rb_series).dt.cast_time_unit(time_unit)
|
4311
|
+
end
|
4312
|
+
s._s
|
4313
|
+
elsif defined?(Numo::NArray) && value.is_a?(Numo::NArray) && value.shape.length == 1
|
4314
|
+
raise Todo
|
4315
|
+
elsif ruby_dtype == ::Array
|
4316
|
+
if dtype.is_a?(Object)
|
4317
|
+
return RbSeries.new_object(name, values, strict)
|
4318
|
+
end
|
4319
|
+
if dtype
|
4320
|
+
srs = sequence_from_anyvalue_or_object(name, values)
|
4321
|
+
if dtype != srs.dtype
|
4322
|
+
srs = srs.cast(dtype, strict: false)
|
4323
|
+
end
|
4324
|
+
return srs
|
4325
|
+
end
|
4326
|
+
sequence_from_anyvalue_or_object(name, values)
|
4327
|
+
elsif ruby_dtype == Series
|
4328
|
+
RbSeries.new_series_list(name, values.map(&:_s), strict)
|
4329
|
+
elsif ruby_dtype == RbSeries
|
4330
|
+
RbSeries.new_series_list(name, values, strict)
|
4331
|
+
else
|
4332
|
+
constructor =
|
4333
|
+
if value.is_a?(::String)
|
4334
|
+
if value.encoding == Encoding::UTF_8
|
4335
|
+
RbSeries.method(:new_str)
|
4336
|
+
else
|
4337
|
+
RbSeries.method(:new_binary)
|
4338
|
+
end
|
4339
|
+
elsif value.is_a?(Integer) && values.any? { |v| v.is_a?(Float) }
|
4340
|
+
# TODO improve performance
|
4341
|
+
RbSeries.method(:new_opt_f64)
|
4342
|
+
else
|
4343
|
+
rb_type_to_constructor(value.class)
|
4344
|
+
end
|
4345
|
+
|
4346
|
+
construct_series_with_fallbacks(constructor, name, values, dtype, strict: strict)
|
4347
|
+
end
|
4348
|
+
end
|
4349
|
+
end
|
4350
|
+
|
4351
|
+
def construct_series_with_fallbacks(constructor, name, values, dtype, strict:)
|
4352
|
+
begin
|
4353
|
+
constructor.call(name, values, strict)
|
4354
|
+
rescue
|
4355
|
+
if dtype.nil?
|
4356
|
+
RbSeries.new_from_any_values(name, values, strict)
|
4357
|
+
else
|
4358
|
+
RbSeries.new_from_any_values_and_dtype(name, values, dtype, strict)
|
4359
|
+
end
|
4360
|
+
end
|
4361
|
+
end
|
4362
|
+
|
4363
|
+
def sequence_from_anyvalue_or_object(name, values)
|
4364
|
+
RbSeries.new_from_any_values(name, values, true)
|
4365
|
+
rescue
|
4366
|
+
RbSeries.new_object(name, values, false)
|
4367
|
+
end
|
4368
|
+
|
4369
|
+
POLARS_TYPE_TO_CONSTRUCTOR = {
|
4370
|
+
Float32 => RbSeries.method(:new_opt_f32),
|
4371
|
+
Float64 => RbSeries.method(:new_opt_f64),
|
4372
|
+
Int8 => RbSeries.method(:new_opt_i8),
|
4373
|
+
Int16 => RbSeries.method(:new_opt_i16),
|
4374
|
+
Int32 => RbSeries.method(:new_opt_i32),
|
4375
|
+
Int64 => RbSeries.method(:new_opt_i64),
|
4376
|
+
UInt8 => RbSeries.method(:new_opt_u8),
|
4377
|
+
UInt16 => RbSeries.method(:new_opt_u16),
|
4378
|
+
UInt32 => RbSeries.method(:new_opt_u32),
|
4379
|
+
UInt64 => RbSeries.method(:new_opt_u64),
|
4380
|
+
Decimal => RbSeries.method(:new_decimal),
|
4381
|
+
Date => RbSeries.method(:new_from_any_values),
|
4382
|
+
Datetime => RbSeries.method(:new_from_any_values),
|
4383
|
+
Duration => RbSeries.method(:new_from_any_values),
|
4384
|
+
Time => RbSeries.method(:new_from_any_values),
|
4385
|
+
Boolean => RbSeries.method(:new_opt_bool),
|
4386
|
+
Utf8 => RbSeries.method(:new_str),
|
4387
|
+
Object => RbSeries.method(:new_object),
|
4388
|
+
Categorical => RbSeries.method(:new_str),
|
4389
|
+
Enum => RbSeries.method(:new_str),
|
4390
|
+
Binary => RbSeries.method(:new_binary),
|
4391
|
+
Null => RbSeries.method(:new_null)
|
4392
|
+
}
|
4393
|
+
|
4394
|
+
SYM_TYPE_TO_CONSTRUCTOR = {
|
4395
|
+
f32: RbSeries.method(:new_opt_f32),
|
4396
|
+
f64: RbSeries.method(:new_opt_f64),
|
4397
|
+
i8: RbSeries.method(:new_opt_i8),
|
4398
|
+
i16: RbSeries.method(:new_opt_i16),
|
4399
|
+
i32: RbSeries.method(:new_opt_i32),
|
4400
|
+
i64: RbSeries.method(:new_opt_i64),
|
4401
|
+
u8: RbSeries.method(:new_opt_u8),
|
4402
|
+
u16: RbSeries.method(:new_opt_u16),
|
4403
|
+
u32: RbSeries.method(:new_opt_u32),
|
4404
|
+
u64: RbSeries.method(:new_opt_u64),
|
4405
|
+
bool: RbSeries.method(:new_opt_bool),
|
4406
|
+
str: RbSeries.method(:new_str)
|
4407
|
+
}
|
4408
|
+
|
4409
|
+
def polars_type_to_constructor(dtype)
|
4410
|
+
if dtype.is_a?(Array)
|
4411
|
+
lambda do |name, values, strict|
|
4412
|
+
RbSeries.new_array(dtype.width, dtype.inner, name, values, strict)
|
4413
|
+
end
|
4414
|
+
elsif dtype.is_a?(Class) && dtype < DataType
|
4415
|
+
POLARS_TYPE_TO_CONSTRUCTOR.fetch(dtype)
|
4416
|
+
elsif dtype.is_a?(DataType)
|
4417
|
+
POLARS_TYPE_TO_CONSTRUCTOR.fetch(dtype.class)
|
4418
|
+
else
|
4419
|
+
SYM_TYPE_TO_CONSTRUCTOR.fetch(dtype.to_sym)
|
4420
|
+
end
|
4421
|
+
rescue KeyError
|
4422
|
+
raise ArgumentError, "Cannot construct RbSeries for type #{dtype}."
|
4423
|
+
end
|
4424
|
+
|
4425
|
+
RB_TYPE_TO_CONSTRUCTOR = {
|
4426
|
+
Float => RbSeries.method(:new_opt_f64),
|
4427
|
+
Integer => RbSeries.method(:new_opt_i64),
|
4428
|
+
TrueClass => RbSeries.method(:new_opt_bool),
|
4429
|
+
FalseClass => RbSeries.method(:new_opt_bool),
|
4430
|
+
BigDecimal => RbSeries.method(:new_decimal),
|
4431
|
+
NilClass => RbSeries.method(:new_null)
|
4432
|
+
}
|
4433
|
+
|
4434
|
+
def rb_type_to_constructor(dtype)
|
4435
|
+
RB_TYPE_TO_CONSTRUCTOR.fetch(dtype)
|
4436
|
+
rescue KeyError
|
4437
|
+
RbSeries.method(:new_object)
|
4438
|
+
end
|
4439
|
+
|
4440
|
+
def _get_first_non_none(values)
|
4441
|
+
values.find { |v| !v.nil? }
|
4442
|
+
end
|
4443
|
+
end
|
4444
|
+
end
|