polars-df 0.7.0-x86_64-darwin → 0.9.0-x86_64-darwin
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +41 -0
- data/Cargo.lock +353 -237
- data/Cargo.toml +0 -3
- data/LICENSE-THIRD-PARTY.txt +4014 -3495
- data/LICENSE.txt +1 -1
- data/README.md +2 -2
- data/lib/polars/3.1/polars.bundle +0 -0
- data/lib/polars/3.2/polars.bundle +0 -0
- data/lib/polars/{3.0 → 3.3}/polars.bundle +0 -0
- data/lib/polars/array_expr.rb +449 -0
- data/lib/polars/array_name_space.rb +346 -0
- data/lib/polars/cat_expr.rb +24 -0
- data/lib/polars/cat_name_space.rb +75 -0
- data/lib/polars/config.rb +2 -2
- data/lib/polars/data_frame.rb +248 -108
- data/lib/polars/data_types.rb +195 -29
- data/lib/polars/date_time_expr.rb +41 -24
- data/lib/polars/date_time_name_space.rb +12 -12
- data/lib/polars/exceptions.rb +12 -1
- data/lib/polars/expr.rb +1080 -195
- 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 +248 -0
- data/lib/polars/functions/col.rb +47 -0
- data/lib/polars/functions/eager.rb +182 -0
- data/lib/polars/functions/lazy.rb +1280 -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 +103 -0
- data/lib/polars/functions/range/int_range.rb +51 -0
- data/lib/polars/functions/repeat.rb +144 -0
- data/lib/polars/functions/whenthen.rb +27 -0
- data/lib/polars/functions.rb +29 -416
- data/lib/polars/group_by.rb +3 -3
- data/lib/polars/io.rb +21 -28
- data/lib/polars/lazy_frame.rb +390 -76
- data/lib/polars/list_expr.rb +152 -6
- data/lib/polars/list_name_space.rb +102 -0
- data/lib/polars/meta_expr.rb +175 -7
- data/lib/polars/series.rb +557 -59
- data/lib/polars/sql_context.rb +1 -1
- data/lib/polars/string_cache.rb +75 -0
- data/lib/polars/string_expr.rb +412 -96
- data/lib/polars/string_name_space.rb +4 -4
- data/lib/polars/struct_expr.rb +1 -1
- data/lib/polars/struct_name_space.rb +1 -1
- data/lib/polars/testing.rb +507 -0
- data/lib/polars/utils.rb +64 -20
- data/lib/polars/version.rb +1 -1
- data/lib/polars.rb +15 -2
- metadata +36 -7
- data/lib/polars/lazy_functions.rb +0 -1197
data/lib/polars/series.rb
CHANGED
@@ -34,7 +34,7 @@ module Polars
|
|
34
34
|
# s3 = Polars::Series.new([1, 2, 3])
|
35
35
|
def initialize(name = nil, values = nil, dtype: nil, strict: true, nan_to_null: false, dtype_if_empty: nil)
|
36
36
|
# Handle case where values are passed as the first argument
|
37
|
-
if !name.nil? && !name.is_a?(String)
|
37
|
+
if !name.nil? && !name.is_a?(::String)
|
38
38
|
if values.nil?
|
39
39
|
values = name
|
40
40
|
name = nil
|
@@ -46,7 +46,7 @@ module Polars
|
|
46
46
|
name = "" if name.nil?
|
47
47
|
|
48
48
|
# TODO improve
|
49
|
-
if values.is_a?(Range) && values.begin.is_a?(String)
|
49
|
+
if values.is_a?(Range) && values.begin.is_a?(::String)
|
50
50
|
values = values.to_a
|
51
51
|
end
|
52
52
|
|
@@ -214,6 +214,126 @@ module Polars
|
|
214
214
|
_comp(other, :lt_eq)
|
215
215
|
end
|
216
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
|
+
|
217
337
|
# Performs addition.
|
218
338
|
#
|
219
339
|
# @return [Series]
|
@@ -341,7 +461,7 @@ module Polars
|
|
341
461
|
def []=(key, value)
|
342
462
|
if value.is_a?(::Array)
|
343
463
|
if is_numeric || is_datelike
|
344
|
-
|
464
|
+
scatter(key, value)
|
345
465
|
return
|
346
466
|
end
|
347
467
|
raise ArgumentError, "cannot set Series of dtype: #{dtype} with list/tuple as value; use a scalar value"
|
@@ -351,9 +471,9 @@ module Polars
|
|
351
471
|
if key.dtype == Boolean
|
352
472
|
self._s = set(key, value)._s
|
353
473
|
elsif key.dtype == UInt64
|
354
|
-
self._s =
|
474
|
+
self._s = scatter(key.cast(UInt32), value)._s
|
355
475
|
elsif key.dtype == UInt32
|
356
|
-
self._s =
|
476
|
+
self._s = scatter(key, value)._s
|
357
477
|
else
|
358
478
|
raise Todo
|
359
479
|
end
|
@@ -411,11 +531,11 @@ module Polars
|
|
411
531
|
# Check if any boolean value in the column is `true`.
|
412
532
|
#
|
413
533
|
# @return [Boolean]
|
414
|
-
def any?(&block)
|
534
|
+
def any?(ignore_nulls: true, &block)
|
415
535
|
if block_given?
|
416
|
-
apply(&block).any?
|
536
|
+
apply(skip_nulls: ignore_nulls, &block).any?
|
417
537
|
else
|
418
|
-
|
538
|
+
_s.any(ignore_nulls)
|
419
539
|
end
|
420
540
|
end
|
421
541
|
alias_method :any, :any?
|
@@ -423,11 +543,11 @@ module Polars
|
|
423
543
|
# Check if all boolean values in the column are `true`.
|
424
544
|
#
|
425
545
|
# @return [Boolean]
|
426
|
-
def all?(&block)
|
546
|
+
def all?(ignore_nulls: true, &block)
|
427
547
|
if block_given?
|
428
|
-
apply(&block).all?
|
548
|
+
apply(skip_nulls: ignore_nulls, &block).all?
|
429
549
|
else
|
430
|
-
|
550
|
+
_s.all(ignore_nulls)
|
431
551
|
end
|
432
552
|
end
|
433
553
|
alias_method :all, :all?
|
@@ -735,6 +855,212 @@ module Polars
|
|
735
855
|
Utils.wrap_df(_s.to_dummies(separator, drop_first))
|
736
856
|
end
|
737
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
|
+
# # │ lengths ┆ values │
|
1022
|
+
# # │ --- ┆ --- │
|
1023
|
+
# # │ i32 ┆ 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
|
+
|
738
1064
|
# Count the unique values in a Series.
|
739
1065
|
#
|
740
1066
|
# @param sort [Boolean]
|
@@ -929,7 +1255,7 @@ module Polars
|
|
929
1255
|
#
|
930
1256
|
# @example
|
931
1257
|
# s = Polars::Series.new("a", [1, 2, 3])
|
932
|
-
# s.
|
1258
|
+
# s.cum_sum
|
933
1259
|
# # =>
|
934
1260
|
# # shape: (3,)
|
935
1261
|
# # Series: 'a' [i64]
|
@@ -938,9 +1264,10 @@ module Polars
|
|
938
1264
|
# # 3
|
939
1265
|
# # 6
|
940
1266
|
# # ]
|
941
|
-
def
|
1267
|
+
def cum_sum(reverse: false)
|
942
1268
|
super
|
943
1269
|
end
|
1270
|
+
alias_method :cumsum, :cum_sum
|
944
1271
|
|
945
1272
|
# Get an array with the cumulative min computed at every element.
|
946
1273
|
#
|
@@ -951,7 +1278,7 @@ module Polars
|
|
951
1278
|
#
|
952
1279
|
# @example
|
953
1280
|
# s = Polars::Series.new("a", [3, 5, 1])
|
954
|
-
# s.
|
1281
|
+
# s.cum_min
|
955
1282
|
# # =>
|
956
1283
|
# # shape: (3,)
|
957
1284
|
# # Series: 'a' [i64]
|
@@ -960,9 +1287,10 @@ module Polars
|
|
960
1287
|
# # 3
|
961
1288
|
# # 1
|
962
1289
|
# # ]
|
963
|
-
def
|
1290
|
+
def cum_min(reverse: false)
|
964
1291
|
super
|
965
1292
|
end
|
1293
|
+
alias_method :cummin, :cum_min
|
966
1294
|
|
967
1295
|
# Get an array with the cumulative max computed at every element.
|
968
1296
|
#
|
@@ -973,7 +1301,7 @@ module Polars
|
|
973
1301
|
#
|
974
1302
|
# @example
|
975
1303
|
# s = Polars::Series.new("a", [3, 5, 1])
|
976
|
-
# s.
|
1304
|
+
# s.cum_max
|
977
1305
|
# # =>
|
978
1306
|
# # shape: (3,)
|
979
1307
|
# # Series: 'a' [i64]
|
@@ -982,9 +1310,10 @@ module Polars
|
|
982
1310
|
# # 5
|
983
1311
|
# # 5
|
984
1312
|
# # ]
|
985
|
-
def
|
1313
|
+
def cum_max(reverse: false)
|
986
1314
|
super
|
987
1315
|
end
|
1316
|
+
alias_method :cummax, :cum_max
|
988
1317
|
|
989
1318
|
# Get an array with the cumulative product computed at every element.
|
990
1319
|
#
|
@@ -999,7 +1328,7 @@ module Polars
|
|
999
1328
|
#
|
1000
1329
|
# @example
|
1001
1330
|
# s = Polars::Series.new("a", [1, 2, 3])
|
1002
|
-
# s.
|
1331
|
+
# s.cum_prod
|
1003
1332
|
# # =>
|
1004
1333
|
# # shape: (3,)
|
1005
1334
|
# # Series: 'a' [i64]
|
@@ -1008,9 +1337,10 @@ module Polars
|
|
1008
1337
|
# # 2
|
1009
1338
|
# # 6
|
1010
1339
|
# # ]
|
1011
|
-
def
|
1340
|
+
def cum_prod(reverse: false)
|
1012
1341
|
super
|
1013
1342
|
end
|
1343
|
+
alias_method :cumprod, :cum_prod
|
1014
1344
|
|
1015
1345
|
# Get the first `n` rows.
|
1016
1346
|
#
|
@@ -1237,26 +1567,56 @@ module Polars
|
|
1237
1567
|
# # 2
|
1238
1568
|
# # 1
|
1239
1569
|
# # ]
|
1240
|
-
def sort(reverse: false, in_place: false)
|
1570
|
+
def sort(reverse: false, nulls_last: false, in_place: false)
|
1241
1571
|
if in_place
|
1242
|
-
self._s = _s.sort(reverse)
|
1572
|
+
self._s = _s.sort(reverse, nulls_last)
|
1243
1573
|
self
|
1244
1574
|
else
|
1245
|
-
Utils.wrap_s(_s.sort(reverse))
|
1575
|
+
Utils.wrap_s(_s.sort(reverse, nulls_last))
|
1246
1576
|
end
|
1247
1577
|
end
|
1248
1578
|
|
1249
1579
|
# Return the `k` largest elements.
|
1250
1580
|
#
|
1251
|
-
#
|
1581
|
+
# @param k [Integer]
|
1582
|
+
# Number of elements to return.
|
1583
|
+
#
|
1584
|
+
# @return [Boolean]
|
1585
|
+
#
|
1586
|
+
# @example
|
1587
|
+
# s = Polars::Series.new("a", [2, 5, 1, 4, 3])
|
1588
|
+
# s.top_k(k: 3)
|
1589
|
+
# # =>
|
1590
|
+
# # shape: (3,)
|
1591
|
+
# # Series: 'a' [i64]
|
1592
|
+
# # [
|
1593
|
+
# # 5
|
1594
|
+
# # 4
|
1595
|
+
# # 3
|
1596
|
+
# # ]
|
1597
|
+
def top_k(k: 5)
|
1598
|
+
super
|
1599
|
+
end
|
1600
|
+
|
1601
|
+
# Return the `k` smallest elements.
|
1252
1602
|
#
|
1253
1603
|
# @param k [Integer]
|
1254
1604
|
# Number of elements to return.
|
1255
|
-
# @param reverse [Boolean]
|
1256
|
-
# Return the smallest elements.
|
1257
1605
|
#
|
1258
1606
|
# @return [Boolean]
|
1259
|
-
|
1607
|
+
#
|
1608
|
+
# @example
|
1609
|
+
# s = Polars::Series.new("a", [2, 5, 1, 4, 3])
|
1610
|
+
# s.bottom_k(k: 3)
|
1611
|
+
# # =>
|
1612
|
+
# # shape: (3,)
|
1613
|
+
# # Series: 'a' [i64]
|
1614
|
+
# # [
|
1615
|
+
# # 1
|
1616
|
+
# # 2
|
1617
|
+
# # 3
|
1618
|
+
# # ]
|
1619
|
+
def bottom_k(k: 5)
|
1260
1620
|
super
|
1261
1621
|
end
|
1262
1622
|
|
@@ -1705,26 +2065,38 @@ module Polars
|
|
1705
2065
|
# @example
|
1706
2066
|
# s = Polars::Series.new("a", [1, 2, 3])
|
1707
2067
|
# s2 = Polars::Series.new("b", [4, 5, 6])
|
1708
|
-
# s.
|
2068
|
+
# s.equals(s)
|
1709
2069
|
# # => true
|
1710
|
-
# s.
|
2070
|
+
# s.equals(s2)
|
1711
2071
|
# # => false
|
1712
|
-
def
|
1713
|
-
_s.
|
2072
|
+
def equals(other, null_equal: false, strict: false)
|
2073
|
+
_s.equals(other._s, null_equal, strict)
|
1714
2074
|
end
|
2075
|
+
alias_method :series_equal, :equals
|
1715
2076
|
|
1716
|
-
#
|
2077
|
+
# Return the number of elements in the Series.
|
1717
2078
|
#
|
1718
2079
|
# @return [Integer]
|
1719
2080
|
#
|
1720
2081
|
# @example
|
1721
|
-
# s = Polars::Series.new("a", [1, 2,
|
2082
|
+
# s = Polars::Series.new("a", [1, 2, nil])
|
2083
|
+
# s.count
|
2084
|
+
# # => 2
|
2085
|
+
def count
|
2086
|
+
len - null_count
|
2087
|
+
end
|
2088
|
+
|
2089
|
+
# Return the number of elements in the Series.
|
2090
|
+
#
|
2091
|
+
# @return [Integer]
|
2092
|
+
#
|
2093
|
+
# @example
|
2094
|
+
# s = Polars::Series.new("a", [1, 2, nil])
|
1722
2095
|
# s.len
|
1723
2096
|
# # => 3
|
1724
2097
|
def len
|
1725
2098
|
_s.len
|
1726
2099
|
end
|
1727
|
-
alias_method :count, :len
|
1728
2100
|
alias_method :length, :len
|
1729
2101
|
alias_method :size, :len
|
1730
2102
|
|
@@ -1886,7 +2258,7 @@ module Polars
|
|
1886
2258
|
# s.is_utf8
|
1887
2259
|
# # => true
|
1888
2260
|
def is_utf8
|
1889
|
-
dtype ==
|
2261
|
+
dtype == String
|
1890
2262
|
end
|
1891
2263
|
alias_method :utf8?, :is_utf8
|
1892
2264
|
|
@@ -1920,7 +2292,7 @@ module Polars
|
|
1920
2292
|
Int64 => Numo::Int64,
|
1921
2293
|
Float32 => Numo::SFloat,
|
1922
2294
|
Float64 => Numo::DFloat
|
1923
|
-
}.fetch(dtype).cast(to_a)
|
2295
|
+
}.fetch(dtype.class).cast(to_a)
|
1924
2296
|
elsif is_boolean
|
1925
2297
|
Numo::Bit.cast(to_a)
|
1926
2298
|
else
|
@@ -1959,7 +2331,7 @@ module Polars
|
|
1959
2331
|
# # 3
|
1960
2332
|
# # ]
|
1961
2333
|
def set(filter, value)
|
1962
|
-
Utils.wrap_s(_s.send("set_with_mask_#{DTYPE_TO_FFINAME.fetch(dtype)}", filter._s, value))
|
2334
|
+
Utils.wrap_s(_s.send("set_with_mask_#{DTYPE_TO_FFINAME.fetch(dtype.class)}", filter._s, value))
|
1963
2335
|
end
|
1964
2336
|
|
1965
2337
|
# Set values at the index locations.
|
@@ -1982,7 +2354,7 @@ module Polars
|
|
1982
2354
|
# # 10
|
1983
2355
|
# # 3
|
1984
2356
|
# # ]
|
1985
|
-
def
|
2357
|
+
def scatter(idx, value)
|
1986
2358
|
if idx.is_a?(Integer)
|
1987
2359
|
idx = [idx]
|
1988
2360
|
end
|
@@ -1991,7 +2363,7 @@ module Polars
|
|
1991
2363
|
end
|
1992
2364
|
|
1993
2365
|
idx = Series.new("", idx)
|
1994
|
-
if value.is_a?(Integer) || value.is_a?(Float) || Utils.bool?(value) || value.is_a?(String) || value.nil?
|
2366
|
+
if value.is_a?(Integer) || value.is_a?(Float) || Utils.bool?(value) || value.is_a?(::String) || value.nil?
|
1995
2367
|
value = Series.new("", [value])
|
1996
2368
|
|
1997
2369
|
# if we need to set more than a single value, we extend it
|
@@ -2001,9 +2373,10 @@ module Polars
|
|
2001
2373
|
elsif !value.is_a?(Series)
|
2002
2374
|
value = Series.new("", value)
|
2003
2375
|
end
|
2004
|
-
_s.
|
2376
|
+
_s.scatter(idx._s, value._s)
|
2005
2377
|
self
|
2006
2378
|
end
|
2379
|
+
alias_method :set_at_idx, :scatter
|
2007
2380
|
|
2008
2381
|
# Create an empty copy of the current Series.
|
2009
2382
|
#
|
@@ -2484,7 +2857,7 @@ module Polars
|
|
2484
2857
|
#
|
2485
2858
|
# @example
|
2486
2859
|
# s = Polars::Series.new("a", [1, 2, 3])
|
2487
|
-
# s.
|
2860
|
+
# s.map_elements { |x| x + 10 }
|
2488
2861
|
# # =>
|
2489
2862
|
# # shape: (3,)
|
2490
2863
|
# # Series: 'a' [i64]
|
@@ -2493,7 +2866,7 @@ module Polars
|
|
2493
2866
|
# # 12
|
2494
2867
|
# # 13
|
2495
2868
|
# # ]
|
2496
|
-
def
|
2869
|
+
def map_elements(return_dtype: nil, skip_nulls: true, &func)
|
2497
2870
|
if return_dtype.nil?
|
2498
2871
|
pl_return_dtype = nil
|
2499
2872
|
else
|
@@ -2501,7 +2874,8 @@ module Polars
|
|
2501
2874
|
end
|
2502
2875
|
Utils.wrap_s(_s.apply_lambda(func, pl_return_dtype, skip_nulls))
|
2503
2876
|
end
|
2504
|
-
alias_method :map, :
|
2877
|
+
alias_method :map, :map_elements
|
2878
|
+
alias_method :apply, :map_elements
|
2505
2879
|
|
2506
2880
|
# Shift the values by a given period.
|
2507
2881
|
#
|
@@ -2830,7 +3204,8 @@ module Polars
|
|
2830
3204
|
weights: nil,
|
2831
3205
|
min_periods: nil,
|
2832
3206
|
center: false,
|
2833
|
-
ddof: 1
|
3207
|
+
ddof: 1,
|
3208
|
+
warn_if_unsorted: true
|
2834
3209
|
)
|
2835
3210
|
to_frame
|
2836
3211
|
.select(
|
@@ -2839,7 +3214,8 @@ module Polars
|
|
2839
3214
|
weights: weights,
|
2840
3215
|
min_periods: min_periods,
|
2841
3216
|
center: center,
|
2842
|
-
ddof: ddof
|
3217
|
+
ddof: ddof,
|
3218
|
+
warn_if_unsorted: warn_if_unsorted
|
2843
3219
|
)
|
2844
3220
|
)
|
2845
3221
|
.to_series
|
@@ -2883,7 +3259,8 @@ module Polars
|
|
2883
3259
|
weights: nil,
|
2884
3260
|
min_periods: nil,
|
2885
3261
|
center: false,
|
2886
|
-
ddof: 1
|
3262
|
+
ddof: 1,
|
3263
|
+
warn_if_unsorted: true
|
2887
3264
|
)
|
2888
3265
|
to_frame
|
2889
3266
|
.select(
|
@@ -2892,7 +3269,8 @@ module Polars
|
|
2892
3269
|
weights: weights,
|
2893
3270
|
min_periods: min_periods,
|
2894
3271
|
center: center,
|
2895
|
-
ddof: ddof
|
3272
|
+
ddof: ddof,
|
3273
|
+
warn_if_unsorted: warn_if_unsorted
|
2896
3274
|
)
|
2897
3275
|
)
|
2898
3276
|
.to_series
|
@@ -2934,7 +3312,8 @@ module Polars
|
|
2934
3312
|
window_size,
|
2935
3313
|
weights: nil,
|
2936
3314
|
min_periods: nil,
|
2937
|
-
center: false
|
3315
|
+
center: false,
|
3316
|
+
warn_if_unsorted: true
|
2938
3317
|
)
|
2939
3318
|
if min_periods.nil?
|
2940
3319
|
min_periods = window_size
|
@@ -2946,7 +3325,8 @@ module Polars
|
|
2946
3325
|
window_size,
|
2947
3326
|
weights: weights,
|
2948
3327
|
min_periods: min_periods,
|
2949
|
-
center: center
|
3328
|
+
center: center,
|
3329
|
+
warn_if_unsorted: warn_if_unsorted
|
2950
3330
|
)
|
2951
3331
|
)
|
2952
3332
|
.to_series
|
@@ -3005,7 +3385,8 @@ module Polars
|
|
3005
3385
|
window_size: 2,
|
3006
3386
|
weights: nil,
|
3007
3387
|
min_periods: nil,
|
3008
|
-
center: false
|
3388
|
+
center: false,
|
3389
|
+
warn_if_unsorted: true
|
3009
3390
|
)
|
3010
3391
|
if min_periods.nil?
|
3011
3392
|
min_periods = window_size
|
@@ -3019,7 +3400,8 @@ module Polars
|
|
3019
3400
|
window_size: window_size,
|
3020
3401
|
weights: weights,
|
3021
3402
|
min_periods: min_periods,
|
3022
|
-
center: center
|
3403
|
+
center: center,
|
3404
|
+
warn_if_unsorted: warn_if_unsorted
|
3023
3405
|
)
|
3024
3406
|
)
|
3025
3407
|
.to_series
|
@@ -3076,8 +3458,8 @@ module Polars
|
|
3076
3458
|
# # shape: (2,)
|
3077
3459
|
# # Series: 'a' [i64]
|
3078
3460
|
# # [
|
3079
|
-
# #
|
3080
|
-
# #
|
3461
|
+
# # 5
|
3462
|
+
# # 3
|
3081
3463
|
# # ]
|
3082
3464
|
def sample(
|
3083
3465
|
n: nil,
|
@@ -3456,6 +3838,113 @@ module Polars
|
|
3456
3838
|
super
|
3457
3839
|
end
|
3458
3840
|
|
3841
|
+
# Replace values by different values.
|
3842
|
+
#
|
3843
|
+
# @param old [Object]
|
3844
|
+
# Value or sequence of values to replace.
|
3845
|
+
# Also accepts a mapping of values to their replacement.
|
3846
|
+
# @param new [Object]
|
3847
|
+
# Value or sequence of values to replace by.
|
3848
|
+
# Length must match the length of `old` or have length 1.
|
3849
|
+
# @param default [Object]
|
3850
|
+
# Set values that were not replaced to this value.
|
3851
|
+
# Defaults to keeping the original value.
|
3852
|
+
# Accepts expression input. Non-expression inputs are parsed as literals.
|
3853
|
+
# @param return_dtype [Object]
|
3854
|
+
# The data type of the resulting Series. If set to `nil` (default),
|
3855
|
+
# the data type is determined automatically based on the other inputs.
|
3856
|
+
#
|
3857
|
+
# @return [Series]
|
3858
|
+
#
|
3859
|
+
# @example Replace a single value by another value. Values that were not replaced remain unchanged.
|
3860
|
+
# s = Polars::Series.new([1, 2, 2, 3])
|
3861
|
+
# s.replace(2, 100)
|
3862
|
+
# # =>
|
3863
|
+
# # shape: (4,)
|
3864
|
+
# # Series: '' [i64]
|
3865
|
+
# # [
|
3866
|
+
# # 1
|
3867
|
+
# # 100
|
3868
|
+
# # 100
|
3869
|
+
# # 3
|
3870
|
+
# # ]
|
3871
|
+
#
|
3872
|
+
# @example Replace multiple values by passing sequences to the `old` and `new` parameters.
|
3873
|
+
# s.replace([2, 3], [100, 200])
|
3874
|
+
# # =>
|
3875
|
+
# # shape: (4,)
|
3876
|
+
# # Series: '' [i64]
|
3877
|
+
# # [
|
3878
|
+
# # 1
|
3879
|
+
# # 100
|
3880
|
+
# # 100
|
3881
|
+
# # 200
|
3882
|
+
# # ]
|
3883
|
+
#
|
3884
|
+
# @example Passing a mapping with replacements is also supported as syntactic sugar. Specify a default to set all values that were not matched.
|
3885
|
+
# mapping = {2 => 100, 3 => 200}
|
3886
|
+
# s.replace(mapping, default: -1)
|
3887
|
+
# # =>
|
3888
|
+
# # shape: (4,)
|
3889
|
+
# # Series: '' [i64]
|
3890
|
+
# # [
|
3891
|
+
# # -1
|
3892
|
+
# # 100
|
3893
|
+
# # 100
|
3894
|
+
# # 200
|
3895
|
+
# # ]
|
3896
|
+
#
|
3897
|
+
# @example The default can be another Series.
|
3898
|
+
# default = Polars::Series.new([2.5, 5.0, 7.5, 10.0])
|
3899
|
+
# s.replace(2, 100, default: default)
|
3900
|
+
# # =>
|
3901
|
+
# # shape: (4,)
|
3902
|
+
# # Series: '' [f64]
|
3903
|
+
# # [
|
3904
|
+
# # 2.5
|
3905
|
+
# # 100.0
|
3906
|
+
# # 100.0
|
3907
|
+
# # 10.0
|
3908
|
+
# # ]
|
3909
|
+
#
|
3910
|
+
# @example Replacing by values of a different data type sets the return type based on a combination of the `new` data type and either the original data type or the default data type if it was set.
|
3911
|
+
# s = Polars::Series.new(["x", "y", "z"])
|
3912
|
+
# mapping = {"x" => 1, "y" => 2, "z" => 3}
|
3913
|
+
# s.replace(mapping)
|
3914
|
+
# # =>
|
3915
|
+
# # shape: (3,)
|
3916
|
+
# # Series: '' [str]
|
3917
|
+
# # [
|
3918
|
+
# # "1"
|
3919
|
+
# # "2"
|
3920
|
+
# # "3"
|
3921
|
+
# # ]
|
3922
|
+
#
|
3923
|
+
# @example
|
3924
|
+
# s.replace(mapping, default: nil)
|
3925
|
+
# # =>
|
3926
|
+
# # shape: (3,)
|
3927
|
+
# # Series: '' [i64]
|
3928
|
+
# # [
|
3929
|
+
# # 1
|
3930
|
+
# # 2
|
3931
|
+
# # 3
|
3932
|
+
# # ]
|
3933
|
+
#
|
3934
|
+
# @example Set the `return_dtype` parameter to control the resulting data type directly.
|
3935
|
+
# s.replace(mapping, return_dtype: Polars::UInt8)
|
3936
|
+
# # =>
|
3937
|
+
# # shape: (3,)
|
3938
|
+
# # Series: '' [u8]
|
3939
|
+
# # [
|
3940
|
+
# # 1
|
3941
|
+
# # 2
|
3942
|
+
# # 3
|
3943
|
+
# # ]
|
3944
|
+
def replace(old, new = Expr::NO_DEFAULT, default: Expr::NO_DEFAULT, return_dtype: nil)
|
3945
|
+
super
|
3946
|
+
end
|
3947
|
+
|
3459
3948
|
# Reshape this Series to a flat Series or a Series of Lists.
|
3460
3949
|
#
|
3461
3950
|
# @param dims [Array]
|
@@ -3665,7 +4154,7 @@ module Polars
|
|
3665
4154
|
end
|
3666
4155
|
|
3667
4156
|
def _pos_idxs(idxs)
|
3668
|
-
idx_type =
|
4157
|
+
idx_type = Plr.get_index_type
|
3669
4158
|
|
3670
4159
|
if idxs.is_a?(Series)
|
3671
4160
|
if idxs.dtype == idx_type
|
@@ -3750,7 +4239,7 @@ module Polars
|
|
3750
4239
|
end
|
3751
4240
|
|
3752
4241
|
def ffi_func(name, dtype, _s)
|
3753
|
-
_s.method(name.sub("<>", DTYPE_TO_FFINAME.fetch(dtype))) if DTYPE_TO_FFINAME.key?(dtype)
|
4242
|
+
_s.method(name.sub("<>", DTYPE_TO_FFINAME.fetch(dtype.class))) if DTYPE_TO_FFINAME.key?(dtype.class)
|
3754
4243
|
end
|
3755
4244
|
|
3756
4245
|
def _arithmetic(other, op)
|
@@ -3761,7 +4250,7 @@ module Polars
|
|
3761
4250
|
return Utils.wrap_s(_s.send(op, other._s))
|
3762
4251
|
end
|
3763
4252
|
|
3764
|
-
if (other.is_a?(Float) || other.is_a?(::Date) || other.is_a?(::DateTime) || other.is_a?(::Time) || other.is_a?(String)) && !is_float
|
4253
|
+
if (other.is_a?(Float) || other.is_a?(::Date) || other.is_a?(::DateTime) || other.is_a?(::Time) || other.is_a?(::String)) && !is_float
|
3765
4254
|
_s2 = sequence_to_rbseries(name, [other])
|
3766
4255
|
return Utils.wrap_s(_s.send(op, _s2))
|
3767
4256
|
end
|
@@ -3865,21 +4354,29 @@ module Polars
|
|
3865
4354
|
end
|
3866
4355
|
end
|
3867
4356
|
|
3868
|
-
if !dtype.nil? && ![List, Unknown].include?(dtype) && Utils.is_polars_dtype(dtype) && ruby_dtype.nil?
|
4357
|
+
if !dtype.nil? && ![List, Struct, Unknown].include?(dtype) && Utils.is_polars_dtype(dtype) && ruby_dtype.nil?
|
3869
4358
|
if dtype == Array && !dtype.is_a?(Array) && value.is_a?(::Array)
|
3870
|
-
dtype = Array.new(value.size)
|
4359
|
+
dtype = Array.new(nil, value.size)
|
3871
4360
|
end
|
3872
4361
|
|
3873
4362
|
constructor = polars_type_to_constructor(dtype)
|
3874
4363
|
rbseries = constructor.call(name, values, strict)
|
3875
4364
|
|
3876
4365
|
base_type = dtype.is_a?(DataType) ? dtype.class : dtype
|
3877
|
-
if [Date, Datetime, Duration, Time, Categorical, Boolean].include?(base_type)
|
4366
|
+
if [Date, Datetime, Duration, Time, Categorical, Boolean, Enum].include?(base_type)
|
3878
4367
|
if rbseries.dtype != dtype
|
3879
4368
|
rbseries = rbseries.cast(dtype, true)
|
3880
4369
|
end
|
3881
4370
|
end
|
3882
4371
|
return rbseries
|
4372
|
+
elsif dtype == Struct
|
4373
|
+
struct_schema = dtype.is_a?(Struct) ? dtype.to_schema : nil
|
4374
|
+
empty = {}
|
4375
|
+
return DataFrame.sequence_to_rbdf(
|
4376
|
+
values.map { |v| v.nil? ? empty : v },
|
4377
|
+
schema: struct_schema,
|
4378
|
+
orient: "row",
|
4379
|
+
).to_struct(name)
|
3883
4380
|
else
|
3884
4381
|
if ruby_dtype.nil?
|
3885
4382
|
if value.nil?
|
@@ -3927,7 +4424,7 @@ module Polars
|
|
3927
4424
|
return RbSeries.new_series_list(name, values, strict)
|
3928
4425
|
else
|
3929
4426
|
constructor =
|
3930
|
-
if value.is_a?(String)
|
4427
|
+
if value.is_a?(::String)
|
3931
4428
|
if value.encoding == Encoding::UTF_8
|
3932
4429
|
RbSeries.method(:new_str)
|
3933
4430
|
else
|
@@ -3970,6 +4467,7 @@ module Polars
|
|
3970
4467
|
Utf8 => RbSeries.method(:new_str),
|
3971
4468
|
Object => RbSeries.method(:new_object),
|
3972
4469
|
Categorical => RbSeries.method(:new_str),
|
4470
|
+
Enum => RbSeries.method(:new_str),
|
3973
4471
|
Binary => RbSeries.method(:new_binary),
|
3974
4472
|
Null => RbSeries.method(:new_null)
|
3975
4473
|
}
|