red-arrow 2.0.0 → 5.0.0

Sign up to get free protection for your applications and to get access to all the features.
Files changed (70) hide show
  1. checksums.yaml +4 -4
  2. data/ext/arrow/arrow.cpp +3 -0
  3. data/ext/arrow/converters.hpp +15 -2
  4. data/ext/arrow/memory-view.cpp +311 -0
  5. data/ext/arrow/memory-view.hpp +26 -0
  6. data/ext/arrow/raw-records.cpp +1 -0
  7. data/ext/arrow/values.cpp +1 -0
  8. data/lib/arrow/array-builder.rb +11 -6
  9. data/lib/arrow/array.rb +130 -0
  10. data/lib/arrow/bigdecimal-extension.rb +5 -1
  11. data/lib/arrow/buffer.rb +10 -6
  12. data/lib/arrow/constructor-arguments-gc-guardable.rb +25 -0
  13. data/lib/arrow/data-type.rb +14 -5
  14. data/lib/arrow/datum.rb +98 -0
  15. data/lib/arrow/decimal128-array-builder.rb +21 -25
  16. data/lib/arrow/decimal128-data-type.rb +2 -0
  17. data/lib/arrow/decimal128.rb +18 -0
  18. data/lib/arrow/decimal256-array-builder.rb +61 -0
  19. data/lib/arrow/decimal256-array.rb +25 -0
  20. data/lib/arrow/decimal256-data-type.rb +73 -0
  21. data/lib/arrow/decimal256.rb +60 -0
  22. data/lib/arrow/dense-union-data-type.rb +2 -2
  23. data/lib/arrow/dictionary-data-type.rb +2 -2
  24. data/lib/arrow/equal-options.rb +38 -0
  25. data/lib/arrow/fixed-size-binary-array-builder.rb +38 -0
  26. data/lib/arrow/fixed-size-binary-array.rb +26 -0
  27. data/lib/arrow/loader.rb +46 -0
  28. data/lib/arrow/scalar.rb +32 -0
  29. data/lib/arrow/sort-key.rb +193 -0
  30. data/lib/arrow/sort-options.rb +109 -0
  31. data/lib/arrow/sparse-union-data-type.rb +2 -2
  32. data/lib/arrow/table.rb +2 -2
  33. data/lib/arrow/time32-data-type.rb +2 -2
  34. data/lib/arrow/time64-data-type.rb +2 -2
  35. data/lib/arrow/timestamp-data-type.rb +2 -2
  36. data/lib/arrow/version.rb +1 -1
  37. data/red-arrow.gemspec +3 -1
  38. data/test/helper.rb +1 -0
  39. data/test/raw-records/test-basic-arrays.rb +17 -0
  40. data/test/raw-records/test-dense-union-array.rb +14 -0
  41. data/test/raw-records/test-list-array.rb +20 -0
  42. data/test/raw-records/test-sparse-union-array.rb +14 -0
  43. data/test/raw-records/test-struct-array.rb +15 -0
  44. data/test/test-array.rb +156 -2
  45. data/test/test-bigdecimal.rb +20 -3
  46. data/test/test-boolean-scalar.rb +26 -0
  47. data/test/test-decimal128-array-builder.rb +18 -1
  48. data/test/test-decimal128-data-type.rb +2 -2
  49. data/test/test-decimal128.rb +38 -0
  50. data/test/test-decimal256-array-builder.rb +112 -0
  51. data/test/test-decimal256-array.rb +38 -0
  52. data/test/test-decimal256-data-type.rb +31 -0
  53. data/test/test-decimal256.rb +102 -0
  54. data/test/test-fixed-size-binary-array-builder.rb +92 -0
  55. data/test/test-fixed-size-binary-array.rb +36 -0
  56. data/test/test-float-scalar.rb +46 -0
  57. data/test/test-function.rb +176 -0
  58. data/test/test-memory-view.rb +434 -0
  59. data/test/test-orc.rb +19 -23
  60. data/test/test-sort-indices.rb +40 -0
  61. data/test/test-sort-key.rb +81 -0
  62. data/test/test-sort-options.rb +58 -0
  63. data/test/test-struct-array-builder.rb +8 -8
  64. data/test/test-struct-array.rb +2 -2
  65. data/test/values/test-basic-arrays.rb +11 -0
  66. data/test/values/test-dense-union-array.rb +14 -0
  67. data/test/values/test-list-array.rb +18 -0
  68. data/test/values/test-sparse-union-array.rb +14 -0
  69. data/test/values/test-struct-array.rb +15 -0
  70. metadata +127 -59
data/test/helper.rb CHANGED
@@ -17,6 +17,7 @@
17
17
 
18
18
  require "arrow"
19
19
 
20
+ require "fiddle"
20
21
  require "pathname"
21
22
  require "tempfile"
22
23
  require "zlib"
@@ -329,6 +329,23 @@ module RawRecordsBasicArraysTests
329
329
  records)
330
330
  assert_equal(records, target.raw_records)
331
331
  end
332
+
333
+ def test_decimal256
334
+ records = [
335
+ [BigDecimal("92.92")],
336
+ [nil],
337
+ [BigDecimal("29.29")],
338
+ ]
339
+ target = build({
340
+ column: {
341
+ type: :decimal256,
342
+ precision: 38,
343
+ scale: 2,
344
+ }
345
+ },
346
+ records)
347
+ assert_equal(records, target.raw_records)
348
+ end
332
349
  end
333
350
 
334
351
  class RawRecordsRecordBatchBasicArraysTest < Test::Unit::TestCase
@@ -345,6 +345,20 @@ module RawRecordsDenseUnionArrayTests
345
345
  assert_equal(records, target.raw_records)
346
346
  end
347
347
 
348
+ def test_decimal256
349
+ records = [
350
+ [{"0" => BigDecimal("92.92")}],
351
+ [{"1" => nil}],
352
+ ]
353
+ target = build({
354
+ type: :decimal256,
355
+ precision: 38,
356
+ scale: 2,
357
+ },
358
+ records)
359
+ assert_equal(records, target.raw_records)
360
+ end
361
+
348
362
  def test_list
349
363
  records = [
350
364
  [{"0" => [true, nil, false]}],
@@ -379,6 +379,26 @@ module RawRecordsListArrayTests
379
379
  assert_equal(records, target.raw_records)
380
380
  end
381
381
 
382
+ def test_decimal256
383
+ records = [
384
+ [
385
+ [
386
+ BigDecimal("92.92"),
387
+ nil,
388
+ BigDecimal("29.29"),
389
+ ],
390
+ ],
391
+ [nil],
392
+ ]
393
+ target = build({
394
+ type: :decimal256,
395
+ precision: 38,
396
+ scale: 2,
397
+ },
398
+ records)
399
+ assert_equal(records, target.raw_records)
400
+ end
401
+
382
402
  def test_list
383
403
  records = [
384
404
  [
@@ -335,6 +335,20 @@ module RawRecordsSparseUnionArrayTests
335
335
  assert_equal(records, target.raw_records)
336
336
  end
337
337
 
338
+ def test_decimal256
339
+ records = [
340
+ [{"0" => BigDecimal("92.92")}],
341
+ [{"1" => nil}],
342
+ ]
343
+ target = build({
344
+ type: :decimal256,
345
+ precision: 38,
346
+ scale: 2,
347
+ },
348
+ records)
349
+ assert_equal(records, target.raw_records)
350
+ end
351
+
338
352
  def test_list
339
353
  records = [
340
354
  [{"0" => [true, nil, false]}],
@@ -329,6 +329,21 @@ module RawRecordsStructArrayTests
329
329
  assert_equal(records, target.raw_records)
330
330
  end
331
331
 
332
+ def test_decimal256
333
+ records = [
334
+ [{"field" => BigDecimal("92.92")}],
335
+ [nil],
336
+ [{"field" => nil}],
337
+ ]
338
+ target = build({
339
+ type: :decimal256,
340
+ precision: 38,
341
+ scale: 2,
342
+ },
343
+ records)
344
+ assert_equal(records, target.raw_records)
345
+ end
346
+
332
347
  def test_list
333
348
  records = [
334
349
  [{"field" => [true, nil, false]}],
data/test/test-array.rb CHANGED
@@ -64,6 +64,40 @@ class ArrayTest < Test::Unit::TestCase
64
64
  end
65
65
  end
66
66
 
67
+ sub_test_case("#equal_array?") do
68
+ test("no options") do
69
+ array1 = Arrow::FloatArray.new([1.1, Float::NAN])
70
+ array2 = Arrow::FloatArray.new([1.1, Float::NAN])
71
+ assert do
72
+ not array1.equal_array?(array2)
73
+ end
74
+ end
75
+
76
+ test("approx") do
77
+ array1 = Arrow::FloatArray.new([1.1])
78
+ array2 = Arrow::FloatArray.new([1.100001])
79
+ assert do
80
+ array1.equal_array?(array2, approx: true)
81
+ end
82
+ end
83
+
84
+ test("nans-equal") do
85
+ array1 = Arrow::FloatArray.new([1.1, Float::NAN])
86
+ array2 = Arrow::FloatArray.new([1.1, Float::NAN])
87
+ assert do
88
+ array1.equal_array?(array2, nans_equal: true)
89
+ end
90
+ end
91
+
92
+ test("absolute-tolerance") do
93
+ array1 = Arrow::FloatArray.new([1.1])
94
+ array2 = Arrow::FloatArray.new([1.101])
95
+ assert do
96
+ array1.equal_array?(array2, approx: true, absolute_tolerance: 0.01)
97
+ end
98
+ end
99
+ end
100
+
67
101
  sub_test_case("#cast") do
68
102
  test("Symbol") do
69
103
  assert_equal(Arrow::Int32Array.new([1, 2, 3]),
@@ -160,12 +194,132 @@ class ArrayTest < Test::Unit::TestCase
160
194
 
161
195
  test("Arrow::ChunkedArray") do
162
196
  chunks = [
163
- Arrow::Int16Array.new([1, 0]),
164
- Arrow::Int16Array.new([1, 0, 3])
197
+ Arrow::Int16Array.new([1, 4]),
198
+ Arrow::Int16Array.new([0, 3])
165
199
  ]
166
200
  right = Arrow::ChunkedArray.new(chunks)
167
201
  assert_equal(Arrow::BooleanArray.new([true, true, true, false]),
168
202
  @array.is_in(right))
169
203
  end
170
204
  end
205
+
206
+ sub_test_case("#concatenate") do
207
+ test("Arrow::Array: same") do
208
+ assert_equal(Arrow::Int32Array.new([1, 2, nil, 4 ,5, 6]),
209
+ Arrow::Int32Array.new([1, 2, nil]).
210
+ concatenate(Arrow::Int32Array.new([4, 5]),
211
+ Arrow::Int32Array.new([6])))
212
+ end
213
+
214
+ test("Arrow::Array: castable") do
215
+ assert_equal(Arrow::Int32Array.new([1, 2, nil, 4 ,5, 6]),
216
+ Arrow::Int32Array.new([1, 2, nil]).
217
+ concatenate(Arrow::Int8Array.new([4, 5]),
218
+ Arrow::UInt32Array.new([6])))
219
+ end
220
+
221
+ test("Arrow::Array: non-castable") do
222
+ assert_raise(Arrow::Error::Invalid) do
223
+ Arrow::Int32Array.new([1, 2, nil]).
224
+ concatenate(Arrow::StringArray.new(["X"]))
225
+ end
226
+ end
227
+
228
+ test("Array") do
229
+ assert_equal(Arrow::Int32Array.new([1, 2, nil, 4 ,nil, 6]),
230
+ Arrow::Int32Array.new([1, 2, nil]).
231
+ concatenate([4, nil],
232
+ [6]))
233
+ end
234
+
235
+ test("invalid") do
236
+ message = "[array][resolve] can't build int32 array: 4"
237
+ assert_raise(ArgumentError.new(message)) do
238
+ Arrow::Int32Array.new([1, 2, nil]).
239
+ concatenate(4)
240
+ end
241
+ end
242
+ end
243
+
244
+ sub_test_case("#+") do
245
+ test("Arrow::Array: same") do
246
+ assert_equal(Arrow::Int32Array.new([1, 2, nil, 4 ,5, 6]),
247
+ Arrow::Int32Array.new([1, 2, nil]) +
248
+ Arrow::Int32Array.new([4, 5, 6]))
249
+ end
250
+
251
+ test("Arrow::Array: castable") do
252
+ assert_equal(Arrow::Int32Array.new([1, 2, nil, 4 ,5, 6]),
253
+ Arrow::Int32Array.new([1, 2, nil]) +
254
+ Arrow::Int8Array.new([4, 5, 6]))
255
+ end
256
+
257
+ test("Arrow::Array: non-castable") do
258
+ assert_raise(Arrow::Error::Invalid) do
259
+ Arrow::Int32Array.new([1, 2, nil]) +
260
+ Arrow::StringArray.new(["X"])
261
+ end
262
+ end
263
+
264
+ test("Array") do
265
+ assert_equal(Arrow::Int32Array.new([1, 2, nil, 4 ,nil, 6]),
266
+ Arrow::Int32Array.new([1, 2, nil]) +
267
+ [4, nil, 6])
268
+ end
269
+
270
+ test("invalid") do
271
+ message = "[array][resolve] can't build int32 array: 4"
272
+ assert_raise(ArgumentError.new(message)) do
273
+ Arrow::Int32Array.new([1, 2, nil]) + 4
274
+ end
275
+ end
276
+ end
277
+
278
+ sub_test_case("#resolve") do
279
+ test("Arrow::Array: same") do
280
+ assert_equal(Arrow::Int32Array.new([1, 2, nil]),
281
+ Arrow::Int32Array.new([]).
282
+ resolve(Arrow::Int32Array.new([1, 2, nil])))
283
+ end
284
+
285
+ test("Arrow::Array: castable") do
286
+ assert_equal(Arrow::Int32Array.new([1, 2, nil]),
287
+ Arrow::Int32Array.new([]).
288
+ resolve(Arrow::Int8Array.new([1, 2, nil])))
289
+ end
290
+
291
+ test("Arrow::Array: non-castable") do
292
+ assert_raise(Arrow::Error::Invalid) do
293
+ Arrow::Int32Array.new([]) +
294
+ Arrow::StringArray.new(["X"])
295
+ end
296
+ end
297
+
298
+ test("Array: non-parametric") do
299
+ assert_equal(Arrow::Int32Array.new([1, 2, nil]),
300
+ Arrow::Int32Array.new([]).
301
+ resolve([1, 2, nil]))
302
+ end
303
+
304
+ test("Array: parametric") do
305
+ list_data_type = Arrow::ListDataType.new(name: "visible", type: :boolean)
306
+ list_array = Arrow::ListArray.new(list_data_type, [])
307
+ assert_equal(Arrow::ListArray.new(list_data_type,
308
+ [
309
+ [true, false],
310
+ nil,
311
+ ]),
312
+ list_array.resolve([
313
+ [true, false],
314
+ nil,
315
+ ]))
316
+ end
317
+
318
+ test("invalid") do
319
+ message = "[array][resolve] can't build int32 array: 4"
320
+ assert_raise(ArgumentError.new(message)) do
321
+ Arrow::Int32Array.new([]).resolve(4)
322
+ end
323
+ end
324
+ end
171
325
  end
@@ -16,8 +16,25 @@
16
16
  # under the License.
17
17
 
18
18
  class BigDecimalTest < Test::Unit::TestCase
19
- test("#to_arrow") do
20
- assert_equal(Arrow::Decimal128.new("3.14"),
21
- BigDecimal("3.14").to_arrow)
19
+ sub_test_case("#to_arrow") do
20
+ def test_128_positive
21
+ assert_equal(Arrow::Decimal128.new("0.1e38"),
22
+ BigDecimal("0.1e38").to_arrow)
23
+ end
24
+
25
+ def test_128_negative
26
+ assert_equal(Arrow::Decimal128.new("-0.1e38"),
27
+ BigDecimal("-0.1e38").to_arrow)
28
+ end
29
+
30
+ def test_256_positive
31
+ assert_equal(Arrow::Decimal256.new("0.1e39"),
32
+ BigDecimal("0.1e39").to_arrow)
33
+ end
34
+
35
+ def test_256_negative
36
+ assert_equal(Arrow::Decimal256.new("-0.1e39"),
37
+ BigDecimal("-0.1e39").to_arrow)
38
+ end
22
39
  end
23
40
  end
@@ -0,0 +1,26 @@
1
+ # Licensed to the Apache Software Foundation (ASF) under one
2
+ # or more contributor license agreements. See the NOTICE file
3
+ # distributed with this work for additional information
4
+ # regarding copyright ownership. The ASF licenses this file
5
+ # to you under the Apache License, Version 2.0 (the
6
+ # "License"); you may not use this file except in compliance
7
+ # with the License. You may obtain a copy of the License at
8
+ #
9
+ # http://www.apache.org/licenses/LICENSE-2.0
10
+ #
11
+ # Unless required by applicable law or agreed to in writing,
12
+ # software distributed under the License is distributed on an
13
+ # "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14
+ # KIND, either express or implied. See the License for the
15
+ # specific language governing permissions and limitations
16
+ # under the License.
17
+
18
+ class BooleanScalarTest < Test::Unit::TestCase
19
+ def setup
20
+ @scalar = Arrow::BooleanScalar.new(true)
21
+ end
22
+
23
+ test("#value") do
24
+ assert_equal(true, @scalar.value)
25
+ end
26
+ end
@@ -80,8 +80,13 @@ class Decimal128ArrayBuilderTest < Test::Unit::TestCase
80
80
  test("is_valids") do
81
81
  @builder.append_values([
82
82
  Arrow::Decimal128.new("10.1"),
83
- nil,
84
83
  Arrow::Decimal128.new("10.1"),
84
+ Arrow::Decimal128.new("10.1"),
85
+ ],
86
+ [
87
+ true,
88
+ false,
89
+ true,
85
90
  ])
86
91
  array = @builder.finish
87
92
  assert_equal([
@@ -91,5 +96,17 @@ class Decimal128ArrayBuilderTest < Test::Unit::TestCase
91
96
  ],
92
97
  array.to_a)
93
98
  end
99
+
100
+ test("packed") do
101
+ @builder.append_values(Arrow::Decimal128.new("10.1").to_bytes.to_s * 3,
102
+ [true, false, true])
103
+ array = @builder.finish
104
+ assert_equal([
105
+ BigDecimal("10.1"),
106
+ nil,
107
+ BigDecimal("10.1"),
108
+ ],
109
+ array.to_a)
110
+ end
94
111
  end
95
112
  end
@@ -18,12 +18,12 @@
18
18
  class Decimal128DataTypeTest < Test::Unit::TestCase
19
19
  sub_test_case(".new") do
20
20
  test("ordered arguments") do
21
- assert_equal("decimal(8, 2)",
21
+ assert_equal("decimal128(8, 2)",
22
22
  Arrow::Decimal128DataType.new(8, 2).to_s)
23
23
  end
24
24
 
25
25
  test("description") do
26
- assert_equal("decimal(8, 2)",
26
+ assert_equal("decimal128(8, 2)",
27
27
  Arrow::Decimal128DataType.new(precision: 8,
28
28
  scale: 2).to_s)
29
29
  end
@@ -60,5 +60,43 @@ class Decimal128Test < Test::Unit::TestCase
60
60
  @decimal128.to_s(1))
61
61
  end
62
62
  end
63
+
64
+ test("#abs") do
65
+ decimal128 = Arrow::Decimal128.new("-10.1")
66
+ assert_equal([
67
+ Arrow::Decimal128.new("-10.1"),
68
+ Arrow::Decimal128.new("10.1"),
69
+ ],
70
+ [
71
+ decimal128,
72
+ decimal128.abs,
73
+ ])
74
+ end
75
+
76
+ test("#abs!") do
77
+ decimal128 = Arrow::Decimal128.new("-10.1")
78
+ decimal128.abs!
79
+ assert_equal(Arrow::Decimal128.new("10.1"),
80
+ decimal128)
81
+ end
82
+
83
+ test("#negate") do
84
+ decimal128 = Arrow::Decimal128.new("-10.1")
85
+ assert_equal([
86
+ Arrow::Decimal128.new("-10.1"),
87
+ Arrow::Decimal128.new("10.1"),
88
+ ],
89
+ [
90
+ decimal128,
91
+ decimal128.negate,
92
+ ])
93
+ end
94
+
95
+ test("#negate!") do
96
+ decimal128 = Arrow::Decimal128.new("-10.1")
97
+ decimal128.negate!
98
+ assert_equal(Arrow::Decimal128.new("10.1"),
99
+ decimal128)
100
+ end
63
101
  end
64
102
  end