polars-df 0.1.3 → 0.1.4
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +5 -0
- data/Cargo.lock +73 -3
- data/Cargo.toml +3 -0
- data/ext/polars/Cargo.toml +12 -1
- data/ext/polars/src/conversion.rs +80 -0
- data/ext/polars/src/error.rs +4 -0
- data/ext/polars/src/lazy/dataframe.rs +2 -2
- data/ext/polars/src/lazy/dsl.rs +98 -0
- data/ext/polars/src/lib.rs +34 -0
- data/ext/polars/src/list_construction.rs +100 -0
- data/ext/polars/src/series.rs +35 -0
- data/lib/polars/cat_name_space.rb +54 -0
- data/lib/polars/data_frame.rb +101 -4
- data/lib/polars/date_time_expr.rb +2 -2
- data/lib/polars/date_time_name_space.rb +1484 -0
- data/lib/polars/expr.rb +3774 -58
- data/lib/polars/expr_dispatch.rb +8 -0
- data/lib/polars/group_by.rb +1 -0
- data/lib/polars/io.rb +1 -1
- data/lib/polars/lazy_frame.rb +8 -4
- data/lib/polars/lazy_functions.rb +126 -16
- data/lib/polars/lazy_group_by.rb +1 -0
- data/lib/polars/list_expr.rb +502 -5
- data/lib/polars/list_name_space.rb +346 -0
- data/lib/polars/series.rb +802 -52
- data/lib/polars/string_expr.rb +189 -13
- data/lib/polars/string_name_space.rb +690 -0
- data/lib/polars/struct_name_space.rb +64 -0
- data/lib/polars/utils.rb +28 -0
- data/lib/polars/version.rb +1 -1
- data/lib/polars.rb +5 -0
- metadata +8 -2
@@ -0,0 +1,346 @@
|
|
1
|
+
module Polars
|
2
|
+
# Series.arr namespace.
|
3
|
+
class ListNameSpace
|
4
|
+
include ExprDispatch
|
5
|
+
|
6
|
+
self._accessor = "arr"
|
7
|
+
|
8
|
+
# @private
|
9
|
+
def initialize(series)
|
10
|
+
self._s = series._s
|
11
|
+
end
|
12
|
+
|
13
|
+
# Get the length of the arrays as UInt32.
|
14
|
+
#
|
15
|
+
# @return [Series]
|
16
|
+
#
|
17
|
+
# @example
|
18
|
+
# s = Polars::Series.new([[1, 2, 3], [5]])
|
19
|
+
# s.arr.lengths
|
20
|
+
# # =>
|
21
|
+
# # shape: (2,)
|
22
|
+
# # Series: '' [u32]
|
23
|
+
# # [
|
24
|
+
# # 3
|
25
|
+
# # 1
|
26
|
+
# # ]
|
27
|
+
def lengths
|
28
|
+
super
|
29
|
+
end
|
30
|
+
|
31
|
+
# Sum all the arrays in the list.
|
32
|
+
#
|
33
|
+
# @return [Series]
|
34
|
+
def sum
|
35
|
+
super
|
36
|
+
end
|
37
|
+
|
38
|
+
# Compute the max value of the arrays in the list.
|
39
|
+
#
|
40
|
+
# @return [Series]
|
41
|
+
def max
|
42
|
+
super
|
43
|
+
end
|
44
|
+
|
45
|
+
# Compute the min value of the arrays in the list.
|
46
|
+
#
|
47
|
+
# @return [Series]
|
48
|
+
def min
|
49
|
+
super
|
50
|
+
end
|
51
|
+
|
52
|
+
# Compute the mean value of the arrays in the list.
|
53
|
+
#
|
54
|
+
# @return [Series]
|
55
|
+
def mean
|
56
|
+
super
|
57
|
+
end
|
58
|
+
|
59
|
+
# Sort the arrays in the list.
|
60
|
+
#
|
61
|
+
# @return [Series]
|
62
|
+
def sort(reverse: false)
|
63
|
+
super
|
64
|
+
end
|
65
|
+
|
66
|
+
# Reverse the arrays in the list.
|
67
|
+
#
|
68
|
+
# @return [Series]
|
69
|
+
def reverse
|
70
|
+
super
|
71
|
+
end
|
72
|
+
|
73
|
+
# Get the unique/distinct values in the list.
|
74
|
+
#
|
75
|
+
# @return [Series]
|
76
|
+
def unique
|
77
|
+
super
|
78
|
+
end
|
79
|
+
|
80
|
+
# Concat the arrays in a Series dtype List in linear time.
|
81
|
+
#
|
82
|
+
# @param other [Object]
|
83
|
+
# Columns to concat into a List Series
|
84
|
+
#
|
85
|
+
# @return [Series]
|
86
|
+
def concat(other)
|
87
|
+
super
|
88
|
+
end
|
89
|
+
|
90
|
+
# Get the value by index in the sublists.
|
91
|
+
#
|
92
|
+
# So index `0` would return the first item of every sublist
|
93
|
+
# and index `-1` would return the last item of every sublist
|
94
|
+
# if an index is out of bounds, it will return a `None`.
|
95
|
+
#
|
96
|
+
# @param index [Integer]
|
97
|
+
# Index to return per sublist
|
98
|
+
#
|
99
|
+
# @return [Series]
|
100
|
+
def get(index)
|
101
|
+
super
|
102
|
+
end
|
103
|
+
|
104
|
+
# Get the value by index in the sublists.
|
105
|
+
#
|
106
|
+
# @return [Series]
|
107
|
+
def [](item)
|
108
|
+
get(item)
|
109
|
+
end
|
110
|
+
|
111
|
+
# Join all string items in a sublist and place a separator between them.
|
112
|
+
#
|
113
|
+
# This errors if inner type of list `!= Utf8`.
|
114
|
+
#
|
115
|
+
# @param separator [String]
|
116
|
+
# string to separate the items with
|
117
|
+
#
|
118
|
+
# @return [Series]
|
119
|
+
#
|
120
|
+
# @example
|
121
|
+
# s = Polars::Series.new([["foo", "bar"], ["hello", "world"]])
|
122
|
+
# s.arr.join("-")
|
123
|
+
# # =>
|
124
|
+
# # shape: (2,)
|
125
|
+
# # Series: '' [str]
|
126
|
+
# # [
|
127
|
+
# # "foo-bar"
|
128
|
+
# # "hello-world"
|
129
|
+
# # ]
|
130
|
+
def join(separator)
|
131
|
+
super
|
132
|
+
end
|
133
|
+
|
134
|
+
# Get the first value of the sublists.
|
135
|
+
#
|
136
|
+
# @return [Series]
|
137
|
+
def first
|
138
|
+
super
|
139
|
+
end
|
140
|
+
|
141
|
+
# Get the last value of the sublists.
|
142
|
+
#
|
143
|
+
# @return [Series]
|
144
|
+
def last
|
145
|
+
super
|
146
|
+
end
|
147
|
+
|
148
|
+
# Check if sublists contain the given item.
|
149
|
+
#
|
150
|
+
# @param item [Object]
|
151
|
+
# Item that will be checked for membership.
|
152
|
+
#
|
153
|
+
# @return [Series]
|
154
|
+
def contains(item)
|
155
|
+
super
|
156
|
+
end
|
157
|
+
|
158
|
+
# Retrieve the index of the minimal value in every sublist.
|
159
|
+
#
|
160
|
+
# @return [Series]
|
161
|
+
def arg_min
|
162
|
+
super
|
163
|
+
end
|
164
|
+
|
165
|
+
# Retrieve the index of the maximum value in every sublist.
|
166
|
+
#
|
167
|
+
# @return [Series]
|
168
|
+
def arg_max
|
169
|
+
super
|
170
|
+
end
|
171
|
+
|
172
|
+
# Calculate the n-th discrete difference of every sublist.
|
173
|
+
#
|
174
|
+
# @param n [Integer]
|
175
|
+
# Number of slots to shift.
|
176
|
+
# @param null_behavior ["ignore", "drop"]
|
177
|
+
# How to handle null values.
|
178
|
+
#
|
179
|
+
# @return [Series]
|
180
|
+
#
|
181
|
+
# @example
|
182
|
+
# s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
|
183
|
+
# s.arr.diff
|
184
|
+
# # =>
|
185
|
+
# # shape: (2,)
|
186
|
+
# # Series: 'a' [list]
|
187
|
+
# # [
|
188
|
+
# # [null, 1, ... 1]
|
189
|
+
# # [null, -8, -1]
|
190
|
+
# # ]
|
191
|
+
def diff(n: 1, null_behavior: "ignore")
|
192
|
+
super
|
193
|
+
end
|
194
|
+
|
195
|
+
# Shift values by the given period.
|
196
|
+
#
|
197
|
+
# @param periods [Integer]
|
198
|
+
# Number of places to shift (may be negative).
|
199
|
+
#
|
200
|
+
# @return [Series]
|
201
|
+
#
|
202
|
+
# @example
|
203
|
+
# s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
|
204
|
+
# s.arr.shift
|
205
|
+
# # =>
|
206
|
+
# # shape: (2,)
|
207
|
+
# # Series: 'a' [list]
|
208
|
+
# # [
|
209
|
+
# # [null, 1, ... 3]
|
210
|
+
# # [null, 10, 2]
|
211
|
+
# # ]
|
212
|
+
def shift(periods = 1)
|
213
|
+
super
|
214
|
+
end
|
215
|
+
|
216
|
+
# Slice every sublist.
|
217
|
+
#
|
218
|
+
# @param offset [Integer]
|
219
|
+
# Start index. Negative indexing is supported.
|
220
|
+
# @param length [Integer]
|
221
|
+
# Length of the slice. If set to `nil` (default), the slice is taken to the
|
222
|
+
# end of the list.
|
223
|
+
#
|
224
|
+
# @return [Series]
|
225
|
+
#
|
226
|
+
# @example
|
227
|
+
# s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
|
228
|
+
# s.arr.slice(1, 2)
|
229
|
+
# # =>
|
230
|
+
# # shape: (2,)
|
231
|
+
# # Series: 'a' [list]
|
232
|
+
# # [
|
233
|
+
# # [2, 3]
|
234
|
+
# # [2, 1]
|
235
|
+
# # ]
|
236
|
+
def slice(offset, length = nil)
|
237
|
+
super
|
238
|
+
end
|
239
|
+
|
240
|
+
# Slice the first `n` values of every sublist.
|
241
|
+
#
|
242
|
+
# @param n [Integer]
|
243
|
+
# Number of values to return for each sublist.
|
244
|
+
#
|
245
|
+
# @return [Series]
|
246
|
+
#
|
247
|
+
# @example
|
248
|
+
# s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
|
249
|
+
# s.arr.head(2)
|
250
|
+
# # =>
|
251
|
+
# # shape: (2,)
|
252
|
+
# # Series: 'a' [list]
|
253
|
+
# # [
|
254
|
+
# # [1, 2]
|
255
|
+
# # [10, 2]
|
256
|
+
# # ]
|
257
|
+
def head(n = 5)
|
258
|
+
super
|
259
|
+
end
|
260
|
+
|
261
|
+
# Slice the last `n` values of every sublist.
|
262
|
+
#
|
263
|
+
# @param n [Integer]
|
264
|
+
# Number of values to return for each sublist.
|
265
|
+
#
|
266
|
+
# @return [Series]
|
267
|
+
#
|
268
|
+
# @example
|
269
|
+
# s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
|
270
|
+
# s.arr.tail(2)
|
271
|
+
# # =>
|
272
|
+
# # shape: (2,)
|
273
|
+
# # Series: 'a' [list]
|
274
|
+
# # [
|
275
|
+
# # [3, 4]
|
276
|
+
# # [2, 1]
|
277
|
+
# # ]
|
278
|
+
def tail(n = 5)
|
279
|
+
super
|
280
|
+
end
|
281
|
+
|
282
|
+
# Convert the series of type `List` to a series of type `Struct`.
|
283
|
+
#
|
284
|
+
# @param n_field_strategy ["first_non_null", "max_width"]
|
285
|
+
# Strategy to determine the number of fields of the struct.
|
286
|
+
# @param name_generator [Object]
|
287
|
+
# A custom function that can be used to generate the field names.
|
288
|
+
# Default field names are `field_0, field_1 .. field_n`
|
289
|
+
#
|
290
|
+
# @return [Series]
|
291
|
+
#
|
292
|
+
# @example
|
293
|
+
# df = Polars::DataFrame.new({"a" => [[1, 2, 3], [1, 2]]})
|
294
|
+
# df.select([Polars.col("a").arr.to_struct])
|
295
|
+
# # =>
|
296
|
+
# # shape: (2, 1)
|
297
|
+
# # ┌────────────┐
|
298
|
+
# # │ a │
|
299
|
+
# # │ --- │
|
300
|
+
# # │ struct[3] │
|
301
|
+
# # ╞════════════╡
|
302
|
+
# # │ {1,2,3} │
|
303
|
+
# # ├╌╌╌╌╌╌╌╌╌╌╌╌┤
|
304
|
+
# # │ {1,2,null} │
|
305
|
+
# # └────────────┘
|
306
|
+
def to_struct(n_field_strategy: "first_non_null", name_generator: nil)
|
307
|
+
super
|
308
|
+
end
|
309
|
+
|
310
|
+
# Run any polars expression against the lists' elements.
|
311
|
+
#
|
312
|
+
# @param expr [Expr]
|
313
|
+
# Expression to run. Note that you can select an element with `Polars.first`, or
|
314
|
+
# `Polars.col`
|
315
|
+
# @param parallel [Boolean]
|
316
|
+
# Run all expression parallel. Don't activate this blindly.
|
317
|
+
# Parallelism is worth it if there is enough work to do per thread.
|
318
|
+
#
|
319
|
+
# This likely should not be use in the groupby context, because we already
|
320
|
+
# parallel execution per group
|
321
|
+
#
|
322
|
+
# @return [Series]
|
323
|
+
#
|
324
|
+
# @example
|
325
|
+
# df = Polars::DataFrame.new({"a" => [1, 8, 3], "b" => [4, 5, 2]})
|
326
|
+
# df.with_column(
|
327
|
+
# Polars.concat_list(["a", "b"]).arr.eval(Polars.element.rank).alias("rank")
|
328
|
+
# )
|
329
|
+
# # =>
|
330
|
+
# # shape: (3, 3)
|
331
|
+
# # ┌─────┬─────┬────────────┐
|
332
|
+
# # │ a ┆ b ┆ rank │
|
333
|
+
# # │ --- ┆ --- ┆ --- │
|
334
|
+
# # │ i64 ┆ i64 ┆ list[f32] │
|
335
|
+
# # ╞═════╪═════╪════════════╡
|
336
|
+
# # │ 1 ┆ 4 ┆ [1.0, 2.0] │
|
337
|
+
# # ├╌╌╌╌╌┼╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌┤
|
338
|
+
# # │ 8 ┆ 5 ┆ [2.0, 1.0] │
|
339
|
+
# # ├╌╌╌╌╌┼╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌┤
|
340
|
+
# # │ 3 ┆ 2 ┆ [2.0, 1.0] │
|
341
|
+
# # └─────┴─────┴────────────┘
|
342
|
+
def eval(expr, parallel: false)
|
343
|
+
super
|
344
|
+
end
|
345
|
+
end
|
346
|
+
end
|