red_amber 0.3.0 → 0.4.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/.rubocop.yml +39 -20
- data/.yardopts +2 -0
- data/CHANGELOG.md +113 -0
- data/Gemfile +1 -1
- data/LICENSE +1 -1
- data/README.md +25 -26
- data/benchmark/basic.yml +2 -2
- data/benchmark/combine.yml +2 -2
- data/benchmark/dataframe.yml +2 -2
- data/benchmark/group.yml +2 -2
- data/benchmark/reshape.yml +2 -2
- data/benchmark/vector.yml +3 -0
- data/doc/DataFrame.md +32 -12
- data/doc/DataFrame_Comparison.md +65 -0
- data/doc/SubFrames.md +11 -0
- data/doc/Vector.md +207 -1
- data/doc/yard-templates/default/fulldoc/html/css/common.css +6 -0
- data/lib/red_amber/data_frame.rb +429 -75
- data/lib/red_amber/data_frame_combinable.rb +516 -66
- data/lib/red_amber/data_frame_displayable.rb +244 -14
- data/lib/red_amber/data_frame_indexable.rb +121 -18
- data/lib/red_amber/data_frame_loadsave.rb +78 -10
- data/lib/red_amber/data_frame_reshaping.rb +184 -14
- data/lib/red_amber/data_frame_selectable.rb +622 -66
- data/lib/red_amber/data_frame_variable_operation.rb +446 -34
- data/lib/red_amber/group.rb +187 -22
- data/lib/red_amber/helper.rb +70 -10
- data/lib/red_amber/refinements.rb +12 -5
- data/lib/red_amber/subframes.rb +1066 -0
- data/lib/red_amber/vector.rb +385 -11
- data/lib/red_amber/vector_aggregation.rb +312 -0
- data/lib/red_amber/vector_binary_element_wise.rb +387 -0
- data/lib/red_amber/vector_selectable.rb +217 -12
- data/lib/red_amber/vector_unary_element_wise.rb +436 -0
- data/lib/red_amber/vector_updatable.rb +278 -34
- data/lib/red_amber/version.rb +2 -1
- data/lib/red_amber.rb +13 -1
- data/red_amber.gemspec +2 -2
- metadata +13 -8
- data/doc/image/dataframe/reshaping_DataFrames.png +0 -0
- data/lib/red_amber/vector_functions.rb +0 -242
@@ -1,39 +1,118 @@
|
|
1
1
|
# frozen_string_literal: true
|
2
2
|
|
3
3
|
module RedAmber
|
4
|
-
#
|
4
|
+
# Mix-in for the class DataFrame
|
5
5
|
module DataFrameVariableOperation
|
6
6
|
# Array is refined
|
7
7
|
using RefineArray
|
8
8
|
|
9
|
-
#
|
9
|
+
# Select variables (columns) to create a new DataFrame.
|
10
10
|
#
|
11
|
-
# @note DataFrame#pick
|
12
|
-
# DataFrame#[]
|
11
|
+
# @note if a single key is specified, DataFrame#pick generates a DataFrame.
|
12
|
+
# On the other hand, DataFrame#[] generates a Vector.
|
13
13
|
#
|
14
14
|
# @overload pick(keys)
|
15
|
-
# Pick variables by
|
15
|
+
# Pick up variables by Symbol(s) or String(s).
|
16
16
|
#
|
17
17
|
# @param keys [Symbol, String, <Symbol, String>]
|
18
18
|
# key name(s) of variables to pick.
|
19
19
|
# @return [DataFrame]
|
20
|
-
#
|
20
|
+
# picked DataFrame.
|
21
|
+
# @example Pick up by a key
|
22
|
+
# languages
|
23
|
+
#
|
24
|
+
# # =>
|
25
|
+
# #<RedAmber::DataFrame : 4 x 3 Vectors, 0x00000000000cfd8c>
|
26
|
+
# Language Creator Released
|
27
|
+
# <string> <string> <uint16>
|
28
|
+
# 0 Ruby Yukihiro Matsumoto 1995
|
29
|
+
# 1 Python Guido van Rossum 1991
|
30
|
+
# 2 R Ross Ihaka and Robert Gentleman 1993
|
31
|
+
# 3 Rust Graydon Hoare 2001
|
32
|
+
#
|
33
|
+
# languages.pick(:Language)
|
34
|
+
#
|
35
|
+
# # =>
|
36
|
+
# #<RedAmber::DataFrame : 4 x 1 Vector, 0x0000000000113d20>
|
37
|
+
# Language
|
38
|
+
# <string>
|
39
|
+
# 0 Ruby
|
40
|
+
# 1 Python
|
41
|
+
# 2 R
|
42
|
+
# 3 Rust
|
43
|
+
#
|
44
|
+
# languages[:Language]
|
45
|
+
#
|
46
|
+
# # =>
|
47
|
+
# #<RedAmber::Vector(:string, size=4):0x000000000010359c>
|
48
|
+
# ["Ruby", "Python", "R", "Rust"]
|
21
49
|
#
|
22
50
|
# @overload pick(booleans)
|
23
|
-
# Pick variables by booleans.
|
51
|
+
# Pick up variables by booleans.
|
24
52
|
#
|
25
|
-
# @param booleans [<
|
26
|
-
# boolean array to pick variables at true.
|
53
|
+
# @param booleans [<Booleans, nil>, Vector]
|
54
|
+
# boolean array or vecctor to pick up variables at true.
|
27
55
|
# @return [DataFrame]
|
28
|
-
#
|
56
|
+
# picked DataFrame.
|
57
|
+
# @example Pick up by booleans
|
58
|
+
# languages.pick(true, true, false)
|
59
|
+
#
|
60
|
+
# # =>
|
61
|
+
# #<RedAmber::DataFrame : 4 x 2 Vectors, 0x0000000000066a1c>
|
62
|
+
# Language Creator
|
63
|
+
# <string> <string>
|
64
|
+
# 0 Ruby Yukihiro Matsumoto
|
65
|
+
# 1 Python Guido van Rossum
|
66
|
+
# 2 R Ross Ihaka and Robert Gentleman
|
67
|
+
# 3 Rust Graydon Hoare
|
68
|
+
#
|
69
|
+
# is_string = languages.vectors.map(&:string?) # [true, true, false]
|
70
|
+
# languages.pick(is_string)
|
71
|
+
# # =>
|
72
|
+
# (same as above)
|
29
73
|
#
|
30
74
|
# @overload pick(indices)
|
31
|
-
# Pick variables by column indices.
|
75
|
+
# Pick up variables by column indices.
|
32
76
|
#
|
33
77
|
# @param indices [Integer, Float, Range<Integer>, Vector, Arrow::Array]
|
34
|
-
# numeric array to pick variables by column index.
|
78
|
+
# numeric array to pick up variables by column index.
|
79
|
+
# @return [DataFrame]
|
80
|
+
# picked DataFrame.
|
81
|
+
# @example Pick up by indices
|
82
|
+
# languages.pick(0, 2, 1)
|
83
|
+
#
|
84
|
+
# # =>
|
85
|
+
# #<RedAmber::DataFrame : 4 x 3 Vectors, 0x000000000011cfb0>
|
86
|
+
# Language Released Creator
|
87
|
+
# <string> <uint16> <string>
|
88
|
+
# 0 Ruby 1995 Yukihiro Matsumoto
|
89
|
+
# 1 Python 1991 Guido van Rossum
|
90
|
+
# 2 R 1993 Ross Ihaka and Robert Gentleman
|
91
|
+
# 3 Rust 2001 Graydon Hoare
|
92
|
+
#
|
93
|
+
# @overload pick
|
94
|
+
# Pick up variables by the yielded value from the block.
|
95
|
+
# @note Arguments and a block cannot be used simultaneously.
|
96
|
+
#
|
97
|
+
# @yield [self]
|
98
|
+
# the block is called within the context of self.
|
99
|
+
# (Block is called by instance_eval(&block). )
|
100
|
+
# @yieldreturn [keys, booleans, indices]
|
101
|
+
# returns keys, booleans or indices just same as arguments.
|
35
102
|
# @return [DataFrame]
|
36
|
-
#
|
103
|
+
# picked DataFrame.
|
104
|
+
# @example Pick up by a block.
|
105
|
+
# # same as languages.pick { |df| df.languages.vectors.map(&:string?) }
|
106
|
+
# languages.pick { languages.vectors.map(&:string?) }
|
107
|
+
#
|
108
|
+
# # =>
|
109
|
+
# #<RedAmber::DataFrame : 4 x 2 Vectors, 0x0000000000154104>
|
110
|
+
# Language Creator
|
111
|
+
# <string> <string>
|
112
|
+
# 0 Ruby Yukihiro Matsumoto
|
113
|
+
# 1 Python Guido van Rossum
|
114
|
+
# 2 R Ross Ihaka and Robert Gentleman
|
115
|
+
# 3 Rust Graydon Hoare
|
37
116
|
#
|
38
117
|
def pick(*args, &block)
|
39
118
|
if block
|
@@ -47,9 +126,9 @@ module RedAmber
|
|
47
126
|
case args
|
48
127
|
in [] | [nil]
|
49
128
|
return DataFrame.new
|
50
|
-
in [*] if args.
|
129
|
+
in [*] if args.symbol?
|
51
130
|
return DataFrame.create(@table.select_columns(*args))
|
52
|
-
in [*] if args.
|
131
|
+
in [*] if args.boolean?
|
53
132
|
picker = keys.select_by_booleans(args)
|
54
133
|
return DataFrame.create(@table.select_columns(*picker))
|
55
134
|
in [(Vector | Arrow::Array | Arrow::ChunkedArray) => a]
|
@@ -60,7 +139,7 @@ module RedAmber
|
|
60
139
|
|
61
140
|
return DataFrame.new if picker.compact.empty?
|
62
141
|
|
63
|
-
if picker.
|
142
|
+
if picker.boolean?
|
64
143
|
picker = keys.select_by_booleans(picker)
|
65
144
|
return DataFrame.create(@table.select_columns(*picker))
|
66
145
|
end
|
@@ -70,33 +149,102 @@ module RedAmber
|
|
70
149
|
DataFrame.create(@table.select_columns(*picker))
|
71
150
|
end
|
72
151
|
|
73
|
-
# Drop some variables (columns) to create a remainer DataFrame
|
152
|
+
# Drop off some variables (columns) to create a remainer DataFrame.
|
74
153
|
#
|
75
|
-
# @note DataFrame#drop creates a DataFrame even if it is a single column
|
154
|
+
# @note DataFrame#drop creates a DataFrame even if it is a single column
|
155
|
+
# (not a Vector).
|
76
156
|
#
|
77
157
|
# @overload drop(keys)
|
78
|
-
# Drop variables by
|
158
|
+
# Drop off variables by Symbol(s) or String(s).
|
79
159
|
#
|
80
160
|
# @param keys [Symbol, String, <Symbol, String>]
|
81
161
|
# key name(s) of variables to drop.
|
82
162
|
# @return [DataFrame]
|
83
|
-
#
|
163
|
+
# remainer DataFrame.
|
164
|
+
# @example Drop off by a key
|
165
|
+
# languages
|
166
|
+
#
|
167
|
+
# # =>
|
168
|
+
# #<RedAmber::DataFrame : 4 x 3 Vectors, 0x00000000000cfd8c>
|
169
|
+
# Language Creator Released
|
170
|
+
# <string> <string> <uint16>
|
171
|
+
# 0 Ruby Yukihiro Matsumoto 1995
|
172
|
+
# 1 Python Guido van Rossum 1991
|
173
|
+
# 2 R Ross Ihaka and Robert Gentleman 1993
|
174
|
+
# 3 Rust Graydon Hoare 2001
|
175
|
+
#
|
176
|
+
# languages.drop(:Language)
|
177
|
+
#
|
178
|
+
# # =>
|
179
|
+
# #<RedAmber::DataFrame : 4 x 2 Vectors, 0x000000000005805c>
|
180
|
+
# Creator Released
|
181
|
+
# <string> <uint16>
|
182
|
+
# 0 Yukihiro Matsumoto 1995
|
183
|
+
# 1 Guido van Rossum 1991
|
184
|
+
# 2 Ross Ihaka and Robert Gentleman 1993
|
185
|
+
# 3 Graydon Hoare 2001
|
84
186
|
#
|
85
187
|
# @overload drop(booleans)
|
86
|
-
# Drop variables by booleans.
|
188
|
+
# Drop off variables by booleans.
|
87
189
|
#
|
88
|
-
# @param booleans [<
|
89
|
-
# boolean array of variables to drop at true.
|
190
|
+
# @param booleans [<Booleans, nil>, Vector]
|
191
|
+
# boolean array or vector of variables to drop at true.
|
90
192
|
# @return [DataFrame]
|
91
|
-
#
|
193
|
+
# remainer DataFrame.
|
194
|
+
# @example Drop off by booleans
|
195
|
+
# is_numeric = languages.vectors.map(&:numeric?) # [nil, nil, true]
|
196
|
+
# languages.drop(is_numeric)
|
197
|
+
#
|
198
|
+
# # =>
|
199
|
+
# #<RedAmber::DataFrame : 4 x 2 Vectors, 0x0000000000066a1c>
|
200
|
+
# Language Creator
|
201
|
+
# <string> <string>
|
202
|
+
# 0 Ruby Yukihiro Matsumoto
|
203
|
+
# 1 Python Guido van Rossum
|
204
|
+
# 2 R Ross Ihaka and Robert Gentleman
|
205
|
+
# 3 Rust Graydon Hoare
|
92
206
|
#
|
93
207
|
# @overload drop(indices)
|
94
|
-
#
|
208
|
+
# Drop off variables by column indices.
|
95
209
|
#
|
96
210
|
# @param indices [Integer, Float, Range<Integer>, Vector, Arrow::Array]
|
97
211
|
# numeric array of variables to drop by column index.
|
98
212
|
# @return [DataFrame]
|
99
|
-
#
|
213
|
+
# remainer DataFrame.
|
214
|
+
# @example Drop off by indices
|
215
|
+
# languages.drop(2)
|
216
|
+
#
|
217
|
+
# # =>
|
218
|
+
# #<RedAmber::DataFrame : 4 x 2 Vectors, 0x0000000000066a1c>
|
219
|
+
# Language Creator
|
220
|
+
# <string> <string>
|
221
|
+
# 0 Ruby Yukihiro Matsumoto
|
222
|
+
# 1 Python Guido van Rossum
|
223
|
+
# 2 R Ross Ihaka and Robert Gentleman
|
224
|
+
# 3 Rust Graydon Hoare
|
225
|
+
#
|
226
|
+
# @overload drop
|
227
|
+
# Drop off variables by the yielded value from the block.
|
228
|
+
# @note Arguments and a block cannot be used simultaneously.
|
229
|
+
#
|
230
|
+
# @yield [self] the block is called within the context of self.
|
231
|
+
# (Block is called by instance_eval(&block). )
|
232
|
+
# @yieldreturn [keys, booleans, indices]
|
233
|
+
# returns keys, booleans or indices just same as arguments.
|
234
|
+
# @return [DataFrame]
|
235
|
+
# remainer DataFrame.
|
236
|
+
# @example Drop off by a block.
|
237
|
+
# # same as languages.drop { |df| df.vectors.map(&:numeric?) }
|
238
|
+
# languages.drop { vectors.map(&:numeric?) }
|
239
|
+
#
|
240
|
+
# # =>
|
241
|
+
# #<RedAmber::DataFrame : 4 x 2 Vectors, 0x0000000000154104>
|
242
|
+
# Language Creator
|
243
|
+
# <string> <string>
|
244
|
+
# 0 Ruby Yukihiro Matsumoto
|
245
|
+
# 1 Python Guido van Rossum
|
246
|
+
# 2 R Ross Ihaka and Robert Gentleman
|
247
|
+
# 3 Rust Graydon Hoare
|
100
248
|
#
|
101
249
|
def drop(*args, &block)
|
102
250
|
if block
|
@@ -109,21 +257,21 @@ module RedAmber
|
|
109
257
|
return self if args.empty? || empty?
|
110
258
|
|
111
259
|
picker =
|
112
|
-
if args.
|
260
|
+
if args.symbol?
|
113
261
|
keys - args
|
114
|
-
elsif args.
|
262
|
+
elsif args.boolean?
|
115
263
|
keys.reject_by_booleans(args)
|
116
|
-
elsif args.
|
264
|
+
elsif args.integer?
|
117
265
|
keys.reject_by_indices(args)
|
118
266
|
else
|
119
267
|
dropper = parse_args(args, n_keys)
|
120
|
-
if dropper.
|
268
|
+
if dropper.boolean?
|
121
269
|
keys.reject_by_booleans(dropper)
|
122
|
-
elsif dropper.
|
270
|
+
elsif dropper.symbol?
|
123
271
|
keys - dropper
|
124
272
|
else
|
125
273
|
dropper.compact!
|
126
|
-
unless dropper.
|
274
|
+
unless dropper.integer?
|
127
275
|
raise DataFrameArgumentError, "Invalid argument #{args}"
|
128
276
|
end
|
129
277
|
|
@@ -136,7 +284,85 @@ module RedAmber
|
|
136
284
|
DataFrame.create(@table.select_columns(*picker))
|
137
285
|
end
|
138
286
|
|
139
|
-
# rename
|
287
|
+
# rename keys (variable/column names) to create a updated DataFrame.
|
288
|
+
#
|
289
|
+
# @overload rename(key_pairs)
|
290
|
+
# Rename by key pairs as a Hash.
|
291
|
+
#
|
292
|
+
# @param key_pairs [Hash{existing_key => new_key}]
|
293
|
+
# key pair(s) of existing name and new name.
|
294
|
+
# @return [DataFrame]
|
295
|
+
# renamed DataFrame.
|
296
|
+
# @example Rename by a Hash
|
297
|
+
# comecome
|
298
|
+
#
|
299
|
+
# # =>
|
300
|
+
# #<RedAmber::DataFrame : 3 x 2 Vectors, 0x00000000000037b4>
|
301
|
+
# name age
|
302
|
+
# <string> <uint8>
|
303
|
+
# 0 Yasuko 68
|
304
|
+
# 1 Rui 49
|
305
|
+
# 2 Hinata 28
|
306
|
+
#
|
307
|
+
# comecome.rename(:age => :age_in_1993)
|
308
|
+
#
|
309
|
+
# # =>
|
310
|
+
# #<RedAmber::DataFrame : 3 x 2 Vectors, 0x00000000000037c8>
|
311
|
+
# name age_in_1993
|
312
|
+
# <string> <uint8>
|
313
|
+
# 0 Yasuko 68
|
314
|
+
# 1 Rui 49
|
315
|
+
# 2 Hinata 28
|
316
|
+
#
|
317
|
+
# @overload rename(key_pairs)
|
318
|
+
# Rename by key pairs as an Array of Array.
|
319
|
+
#
|
320
|
+
# @param key_pairs [<Array[existing_key, new_key]>]
|
321
|
+
# key pair(s) of existing name and new name.
|
322
|
+
# @return [DataFrame]
|
323
|
+
# renamed DataFrame.
|
324
|
+
# @example Rename by an Array
|
325
|
+
# renamer = [[:name, :heroine], [:age, :age_in_1993]]
|
326
|
+
# comecome.rename(renamer)
|
327
|
+
#
|
328
|
+
# # =>
|
329
|
+
# #<RedAmber::DataFrame : 3 x 2 Vectors, 0x00000000000037dc>
|
330
|
+
# heroine age_in_1993
|
331
|
+
# <string> <uint8>
|
332
|
+
# 0 Yasuko 68
|
333
|
+
# 1 Rui 49
|
334
|
+
# 2 Hinata 28
|
335
|
+
#
|
336
|
+
# @overload rename
|
337
|
+
# Rename by key pairs yielding from block.
|
338
|
+
#
|
339
|
+
# @yield [self] the block is called within the context of self.
|
340
|
+
# (Block is called by instance_eval(&block). )
|
341
|
+
# @yieldreturn [<[existing_key, new_key]>, Hash]
|
342
|
+
# returns an Array or a Hash just same as arguments.
|
343
|
+
# @return [DataFrame]
|
344
|
+
# renamed DataFrame.
|
345
|
+
# @example Rename by block.
|
346
|
+
# df
|
347
|
+
#
|
348
|
+
# # =>
|
349
|
+
# #<RedAmber::DataFrame : 2 x 3 Vectors, 0x000000000000c29c>
|
350
|
+
# X Y Z
|
351
|
+
# <uint8> <uint8> <uint8>
|
352
|
+
# 0 1 3 5
|
353
|
+
# 1 2 4 6
|
354
|
+
#
|
355
|
+
# df.rename { keys.zip(keys.map(&:downcase)) }
|
356
|
+
# # or
|
357
|
+
# df.rename { [keys, keys.map(&:downcase)].transpose }
|
358
|
+
#
|
359
|
+
# # =>
|
360
|
+
# #<RedAmber::DataFrame : 2 x 3 Vectors, 0x000000000000c364>
|
361
|
+
# x y z
|
362
|
+
# <uint8> <uint8> <uint8>
|
363
|
+
# 0 1 3 5
|
364
|
+
# 1 2 4 6
|
365
|
+
#
|
140
366
|
def rename(*renamer, &block)
|
141
367
|
if block
|
142
368
|
unless renamer.empty?
|
@@ -162,11 +388,197 @@ module RedAmber
|
|
162
388
|
rename_by_hash(key_pairs)
|
163
389
|
end
|
164
390
|
|
165
|
-
#
|
391
|
+
# Assign new or updated variables (columns) and create an updated DataFrame.
|
392
|
+
# - Array-like variables with new keys will append new columns from right.
|
393
|
+
# - Array-like variables with exisiting keys will update corresponding vectors.
|
394
|
+
# - Symbol key and String key are considered as the same key.
|
395
|
+
# - If assigner is empty or nil, returns self.
|
396
|
+
#
|
397
|
+
# @overload assign(key_value_pairs)
|
398
|
+
# accepts pairs of key and values by an Array or a Hash.
|
399
|
+
#
|
400
|
+
# @param key_value_pairs [Array<key, array_like>, Hash{key => array_like}]
|
401
|
+
# `key` must be a Symbol or a String.
|
402
|
+
# `array_like` is column data to be assigned.
|
403
|
+
# It must be one of `Vector` or `Arrow::Array` or `Array`.
|
404
|
+
# @return [DataFrame]
|
405
|
+
# assigned DataFrame.
|
406
|
+
# @example Assign a new column
|
407
|
+
# comecome
|
408
|
+
#
|
409
|
+
# # =>
|
410
|
+
# #<RedAmber::DataFrame : 3 x 2 Vectors, 0x00000000000280dc>
|
411
|
+
# name age
|
412
|
+
# <string> <uint8>
|
413
|
+
# 0 Yasuko 68
|
414
|
+
# 1 Rui 49
|
415
|
+
# 2 Hinata 28
|
416
|
+
#
|
417
|
+
# brothers = ['Santa', nil, 'Momotaro']
|
418
|
+
# comecome.assign(brother: brothers)
|
419
|
+
# # or
|
420
|
+
# comecome.assign({ brother: brothers })
|
421
|
+
# # or
|
422
|
+
# comecome.assign(:brother, brothers)
|
423
|
+
# # or
|
424
|
+
# comecome.assign([:brother, brothers])
|
425
|
+
#
|
426
|
+
# # =>
|
427
|
+
# #<RedAmber::DataFrame : 3 x 3 Vectors, 0x000000000004077c>
|
428
|
+
# name age brother
|
429
|
+
# <string> <uint8> <string>
|
430
|
+
# 0 Yasuko 68 Santa
|
431
|
+
# 1 Rui 49 (nil)
|
432
|
+
# 2 Hinata 28 Momotaro
|
433
|
+
#
|
434
|
+
# @example Assign new data for a existing column
|
435
|
+
# comecome.assign(age: comecome[:age] + 29)
|
436
|
+
#
|
437
|
+
# # =>
|
438
|
+
# #<RedAmber::DataFrame : 3 x 2 Vectors, 0x0000000000065860>
|
439
|
+
# name age
|
440
|
+
# <string> <uint8>
|
441
|
+
# 0 Yasuko 97
|
442
|
+
# 1 Rui 78
|
443
|
+
# 2 Hinata 57
|
444
|
+
#
|
445
|
+
# @overload assign
|
446
|
+
# accepts block yielding pairs of key and values.
|
447
|
+
#
|
448
|
+
# @yield [self]
|
449
|
+
# the block is called within the context of self.
|
450
|
+
# (Block is called by instance_eval(&block). )
|
451
|
+
# @yieldreturn [Array<key, array_like>, Hash(key => array_like)]
|
452
|
+
# `key` must be a Symbol or a String.
|
453
|
+
# `array_like` is column data to be assigned.
|
454
|
+
# It must be one of `Vector` or `Arrow::Array` or `Array`.
|
455
|
+
# @return [DataFrame]
|
456
|
+
# assigned DataFrame.
|
457
|
+
# @example Assign new data for a existing column by block
|
458
|
+
# comecome.assign { { age: age + 29 } }
|
459
|
+
# # or
|
460
|
+
# comecome.assign { [:age, age + 29] }
|
461
|
+
# # or
|
462
|
+
# comecome.assign { [[:age, age + 29]] }
|
463
|
+
#
|
464
|
+
# # =>
|
465
|
+
# #<RedAmber::DataFrame : 3 x 2 Vectors, 0x000000000007d640>
|
466
|
+
# name age
|
467
|
+
# <string> <uint8>
|
468
|
+
# 0 Yasuko 97
|
469
|
+
# 1 Rui 78
|
470
|
+
# 2 Hinata 57
|
471
|
+
#
|
472
|
+
# @overload assign(keys)
|
473
|
+
# accepts keys from argument and pairs of key and values from block.
|
474
|
+
#
|
475
|
+
# @param keys [Symbol, String] keys of columns to create or update.
|
476
|
+
# @yield [self]
|
477
|
+
# the block is called within the context of self.
|
478
|
+
# (Block is called by instance_eval(&block).)
|
479
|
+
# @yieldreturn [Array<array_like>]
|
480
|
+
# column data to be assigned.
|
481
|
+
# `array_like` must be one of `Vector` or `Arrow::Array` or `Array`.
|
482
|
+
# @return [DataFrame]
|
483
|
+
# assigned DataFrame.
|
484
|
+
# @example Assign new data for a existing column by block
|
485
|
+
# comecome.assign(:age) { age + 29 }
|
486
|
+
#
|
487
|
+
# # =>
|
488
|
+
# #<RedAmber::DataFrame : 3 x 2 Vectors, 0x000000000007af94>
|
489
|
+
# name age
|
490
|
+
# <string> <uint8>
|
491
|
+
# 0 Yasuko 97
|
492
|
+
# 1 Rui 78
|
493
|
+
# 2 Hinata 57
|
494
|
+
#
|
495
|
+
# @example Assign multiple data
|
496
|
+
# comecome.assign(:age_in_1993, :brother) do
|
497
|
+
# [
|
498
|
+
# age + 29,
|
499
|
+
# ['Santa', nil, 'Momotaro'],
|
500
|
+
# ]
|
501
|
+
# end
|
502
|
+
#
|
503
|
+
# # =>
|
504
|
+
# #<RedAmber::DataFrame : 3 x 4 Vectors, 0x00000000000b363c>
|
505
|
+
# name age age_in_1993 brother
|
506
|
+
# <string> <uint8> <uint8> <string>
|
507
|
+
# 0 Yasuko 68 97 Santa
|
508
|
+
# 1 Rui 49 78 (nil)
|
509
|
+
# 2 Hinata 28 57 Momotaro
|
510
|
+
#
|
166
511
|
def assign(*assigner, &block)
|
167
512
|
assign_update(*assigner, append_to_left: false, &block)
|
168
513
|
end
|
169
514
|
|
515
|
+
# Assign new or updated variables (columns) and create an updated DataFrame.
|
516
|
+
# - Array-like variables with new keys will append new columns from left.
|
517
|
+
# - Array-like variables with exisiting keys will update corresponding vectors.
|
518
|
+
# - Symbol key and String key are considered as the same key.
|
519
|
+
# - If assigner is empty or nil, returns self.
|
520
|
+
#
|
521
|
+
# @overload assign_left(key_value_pairs)
|
522
|
+
# accepts pairs of key and values by an Array or a Hash.
|
523
|
+
#
|
524
|
+
# @param key_value_pairs [Array<key, array_like>, Hash{key => array_like}]
|
525
|
+
# `key` must be a Symbol or a String.
|
526
|
+
# `array_like` is column data to be assigned.
|
527
|
+
# It must be one of `Vector` or `Arrow::Array` or `Array`.
|
528
|
+
# @return [DataFrame]
|
529
|
+
# assigned DataFrame.
|
530
|
+
# @example Assign a new column from left
|
531
|
+
# df
|
532
|
+
#
|
533
|
+
# # =>
|
534
|
+
# #<RedAmber::DataFrame : 5 x 3 Vectors, 0x000000000000c10c>
|
535
|
+
# index float string
|
536
|
+
# <uint8> <double> <string>
|
537
|
+
# 0 0 0.0 A
|
538
|
+
# 1 1 1.1 B
|
539
|
+
# 2 2 2.2 C
|
540
|
+
# 3 3 NaN D
|
541
|
+
# 4 (nil) (nil) (nil)
|
542
|
+
#
|
543
|
+
# df.assign_left(new_index: df.indices(1))
|
544
|
+
#
|
545
|
+
# # =>
|
546
|
+
# #<RedAmber::DataFrame : 5 x 4 Vectors, 0x000000000001787c>
|
547
|
+
# new_index index float string
|
548
|
+
# <uint8> <uint8> <double> <string>
|
549
|
+
# 0 1 0 0.0 A
|
550
|
+
# 1 2 1 1.1 B
|
551
|
+
# 2 3 2 2.2 C
|
552
|
+
# 3 4 3 NaN D
|
553
|
+
# 4 5 (nil) (nil) (nil)
|
554
|
+
#
|
555
|
+
# @overload assign_left
|
556
|
+
# accepts block yielding pairs of key and values.
|
557
|
+
#
|
558
|
+
# @yield [self]
|
559
|
+
# the block is called within the context of self.
|
560
|
+
# (Block is called by instance_eval(&block). )
|
561
|
+
# @yieldreturn [Array<key, array_like>, Hash(key => array_like)]
|
562
|
+
# `key` must be a Symbol or a String.
|
563
|
+
# `array_like` is column data to be assigned.
|
564
|
+
# It must be one of `Vector` or `Arrow::Array` or `Array`.
|
565
|
+
# @return [DataFrame]
|
566
|
+
# assigned DataFrame.
|
567
|
+
#
|
568
|
+
# @overload assign_left(keys)
|
569
|
+
# accepts keys from argument and pairs of key and values from block.
|
570
|
+
#
|
571
|
+
# @param keys [Symbol, String]
|
572
|
+
# keys of columns to create or update.
|
573
|
+
# @yield [self]
|
574
|
+
# the block is called within the context of self.
|
575
|
+
# (Block is called by instance_eval(&block).)
|
576
|
+
# @yieldreturn [Array<array_like>]
|
577
|
+
# column data to be assigned.
|
578
|
+
# `array_like` must be one of `Vector` or `Arrow::Array` or `Array`.
|
579
|
+
# @return [DataFrame]
|
580
|
+
# assigned DataFrame.
|
581
|
+
#
|
170
582
|
def assign_left(*assigner, &block)
|
171
583
|
assign_update(*assigner, append_to_left: true, &block)
|
172
584
|
end
|