red-arrow 0.17.1 → 4.0.0

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.
Files changed (72) hide show
  1. checksums.yaml +4 -4
  2. data/ext/arrow/converters.hpp +75 -32
  3. data/ext/arrow/extconf.rb +14 -3
  4. data/ext/arrow/raw-records.cpp +3 -1
  5. data/ext/arrow/values.cpp +3 -1
  6. data/lib/arrow/array-builder.rb +11 -6
  7. data/lib/arrow/array.rb +118 -0
  8. data/lib/arrow/bigdecimal-extension.rb +5 -1
  9. data/lib/arrow/buffer.rb +28 -0
  10. data/lib/arrow/data-type.rb +14 -5
  11. data/lib/arrow/decimal128-array-builder.rb +21 -25
  12. data/lib/arrow/decimal128-data-type.rb +2 -0
  13. data/lib/arrow/decimal128.rb +18 -0
  14. data/lib/arrow/decimal256-array-builder.rb +61 -0
  15. data/lib/arrow/decimal256-array.rb +25 -0
  16. data/lib/arrow/decimal256-data-type.rb +73 -0
  17. data/lib/arrow/decimal256.rb +60 -0
  18. data/lib/arrow/dense-union-data-type.rb +2 -2
  19. data/lib/arrow/dictionary-array.rb +24 -0
  20. data/lib/arrow/dictionary-data-type.rb +2 -2
  21. data/lib/arrow/fixed-size-binary-array-builder.rb +38 -0
  22. data/lib/arrow/fixed-size-binary-array.rb +26 -0
  23. data/lib/arrow/loader.rb +18 -0
  24. data/lib/arrow/raw-table-converter.rb +47 -0
  25. data/lib/arrow/record-batch-iterator.rb +22 -0
  26. data/lib/arrow/record-batch.rb +9 -1
  27. data/lib/arrow/sort-key.rb +193 -0
  28. data/lib/arrow/sort-options.rb +109 -0
  29. data/lib/arrow/sparse-union-data-type.rb +2 -2
  30. data/lib/arrow/struct-array-builder.rb +13 -7
  31. data/lib/arrow/table-saver.rb +6 -6
  32. data/lib/arrow/table.rb +5 -24
  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 +1 -0
  38. data/test/raw-records/test-basic-arrays.rb +17 -0
  39. data/test/raw-records/test-dense-union-array.rb +15 -34
  40. data/test/raw-records/test-list-array.rb +20 -0
  41. data/test/raw-records/test-sparse-union-array.rb +15 -33
  42. data/test/raw-records/test-struct-array.rb +15 -0
  43. data/test/test-array.rb +122 -2
  44. data/test/test-bigdecimal.rb +20 -3
  45. data/test/test-buffer.rb +11 -0
  46. data/test/test-decimal128-array-builder.rb +18 -1
  47. data/test/test-decimal128-data-type.rb +2 -2
  48. data/test/test-decimal128.rb +38 -0
  49. data/test/test-decimal256-array-builder.rb +112 -0
  50. data/test/test-decimal256-array.rb +38 -0
  51. data/test/test-decimal256-data-type.rb +31 -0
  52. data/test/test-decimal256.rb +102 -0
  53. data/test/test-dense-union-data-type.rb +2 -2
  54. data/test/test-dictionary-array.rb +41 -0
  55. data/test/test-feather.rb +1 -1
  56. data/test/test-fixed-size-binary-array-builder.rb +92 -0
  57. data/test/test-fixed-size-binary-array.rb +36 -0
  58. data/test/test-orc.rb +19 -23
  59. data/test/test-record-batch-iterator.rb +37 -0
  60. data/test/test-record-batch.rb +14 -0
  61. data/test/test-sort-indices.rb +40 -0
  62. data/test/test-sort-key.rb +81 -0
  63. data/test/test-sort-options.rb +58 -0
  64. data/test/test-sparse-union-data-type.rb +2 -2
  65. data/test/test-struct-array-builder.rb +16 -12
  66. data/test/test-struct-array.rb +2 -2
  67. data/test/values/test-basic-arrays.rb +11 -0
  68. data/test/values/test-dense-union-array.rb +15 -34
  69. data/test/values/test-list-array.rb +18 -0
  70. data/test/values/test-sparse-union-array.rb +15 -33
  71. data/test/values/test-struct-array.rb +15 -0
  72. metadata +107 -59
data/test/test-feather.rb CHANGED
@@ -28,7 +28,7 @@ class FeatherTest < Test::Unit::TestCase
28
28
  @output = Tempfile.new(["red-arrow", ".feather"])
29
29
  begin
30
30
  yield(@output)
31
- rescue
31
+ ensure
32
32
  @output.close!
33
33
  end
34
34
  end
@@ -0,0 +1,92 @@
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 FixedSizeBinaryArrayBuilderTest < Test::Unit::TestCase
19
+ def setup
20
+ @data_type = Arrow::FixedSizeBinaryDataType.new(4)
21
+ @builder = Arrow::FixedSizeBinaryArrayBuilder.new(@data_type)
22
+ end
23
+
24
+ sub_test_case("#append_value") do
25
+ test("nil") do
26
+ @builder.append_value(nil)
27
+ array = @builder.finish
28
+ assert_equal(nil, array[0])
29
+ end
30
+
31
+ test("String") do
32
+ @builder.append_value("0123")
33
+ array = @builder.finish
34
+ assert_equal("0123", array[0])
35
+ end
36
+
37
+ test("GLib::Bytes") do
38
+ @builder.append_value(GLib::Bytes.new("0123"))
39
+ array = @builder.finish
40
+ assert_equal("0123", array[0])
41
+ end
42
+ end
43
+
44
+ sub_test_case("#append_values") do
45
+ test("mixed") do
46
+ @builder.append_values([
47
+ "0123",
48
+ nil,
49
+ GLib::Bytes.new("abcd"),
50
+ ])
51
+ array = @builder.finish
52
+ assert_equal([
53
+ "0123",
54
+ nil,
55
+ "abcd",
56
+ ],
57
+ array.to_a)
58
+ end
59
+
60
+ test("is_valids") do
61
+ @builder.append_values([
62
+ "0123",
63
+ "0123",
64
+ "0123",
65
+ ],
66
+ [
67
+ true,
68
+ false,
69
+ true,
70
+ ])
71
+ array = @builder.finish
72
+ assert_equal([
73
+ "0123",
74
+ nil,
75
+ "0123",
76
+ ],
77
+ array.to_a)
78
+ end
79
+
80
+ test("packed") do
81
+ @builder.append_values("0123" * 3,
82
+ [true, false, true])
83
+ array = @builder.finish
84
+ assert_equal([
85
+ "0123",
86
+ nil,
87
+ "0123",
88
+ ],
89
+ array.to_a)
90
+ end
91
+ end
92
+ end
@@ -0,0 +1,36 @@
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 FixedSizeBinaryArrayTest < Test::Unit::TestCase
19
+ sub_test_case(".new") do
20
+ test("build") do
21
+ data_type = Arrow::FixedSizeBinaryDataType.new(4)
22
+ values = [
23
+ "0123",
24
+ nil,
25
+ GLib::Bytes.new("abcd"),
26
+ ]
27
+ array = Arrow::FixedSizeBinaryArray.new(data_type, values)
28
+ assert_equal([
29
+ "0123",
30
+ nil,
31
+ "abcd",
32
+ ],
33
+ array.to_a)
34
+ end
35
+ end
36
+ end
data/test/test-orc.rb CHANGED
@@ -118,39 +118,35 @@ class ORCTest < Test::Unit::TestCase
118
118
  ]
119
119
  ],
120
120
  [
121
- "map: list<item: " +
122
- "struct<key: string, value: " +
123
- "struct<int1: int32, string1: string>>>",
121
+ "map: map<string, struct<int1: int32, string1: string>>",
124
122
  [
125
123
  <<-MAP.chomp
126
124
  [
125
+ keys:
126
+ []
127
+ values:
127
128
  -- is_valid: all not null
128
- -- child 0 type: string
129
+ -- child 0 type: int32
129
130
  []
130
- -- child 1 type: struct<int1: int32, string1: string>
131
- -- is_valid: all not null
132
- -- child 0 type: int32
133
- []
134
- -- child 1 type: string
135
- [],
131
+ -- child 1 type: string
132
+ [],
133
+ keys:
134
+ [
135
+ "chani",
136
+ "mauddib"
137
+ ]
138
+ values:
136
139
  -- is_valid: all not null
137
- -- child 0 type: string
140
+ -- child 0 type: int32
141
+ [
142
+ 5,
143
+ 1
144
+ ]
145
+ -- child 1 type: string
138
146
  [
139
147
  "chani",
140
148
  "mauddib"
141
149
  ]
142
- -- child 1 type: struct<int1: int32, string1: string>
143
- -- is_valid: all not null
144
- -- child 0 type: int32
145
- [
146
- 5,
147
- 1
148
- ]
149
- -- child 1 type: string
150
- [
151
- "chani",
152
- "mauddib"
153
- ]
154
150
  ]
155
151
  MAP
156
152
  ],
@@ -0,0 +1,37 @@
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 RecordBatchIteratorTest < Test::Unit::TestCase
19
+ def setup
20
+ @schema = Arrow::Schema.new(visible: :boolean,
21
+ count: :uint32)
22
+ @record_batches = [
23
+ Arrow::RecordBatch.new(@schema,
24
+ visible: [true],
25
+ count: [1]),
26
+ Arrow::RecordBatch.new(@schema,
27
+ visible: [false, nil],
28
+ count: [nil, 3]),
29
+ ]
30
+ @iterator = Arrow::RecordBatchIterator.new(@record_batches)
31
+ end
32
+
33
+ def test_to_a
34
+ assert_equal(@record_batches,
35
+ @iterator.to_a)
36
+ end
37
+ end
@@ -22,6 +22,20 @@ class RecordBatchTest < Test::Unit::TestCase
22
22
  count: :uint32)
23
23
  end
24
24
 
25
+ test("[raw_table]") do
26
+ raw_table = {
27
+ visible: [true, nil, false],
28
+ count: [1, nil, 3],
29
+ }
30
+ record_batch = Arrow::RecordBatch.new(raw_table)
31
+ assert_equal([
32
+ {"visible" => true, "count" => 1},
33
+ {"visible" => nil, "count" => nil},
34
+ {"visible" => false, "count" => 3},
35
+ ],
36
+ record_batch.each_record.collect(&:to_h))
37
+ end
38
+
25
39
  test("[Schema, records]") do
26
40
  records = [
27
41
  {visible: true, count: 1},
@@ -0,0 +1,40 @@
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 SortIndicesTest < Test::Unit::TestCase
19
+ def setup
20
+ @table = Arrow::Table.new(number1: [16, -1, 2, 32, -4, -4, -8],
21
+ number2: [32, 2, -16, 8, 1, 4, 1])
22
+ end
23
+
24
+ sub_test_case("Table") do
25
+ test("Symbol") do
26
+ assert_equal(Arrow::UInt64Array.new([6, 4, 5, 1, 2, 0, 3]),
27
+ @table.sort_indices(:number1))
28
+ end
29
+
30
+ test("-String") do
31
+ assert_equal(Arrow::UInt64Array.new([3, 0, 2, 1, 4, 5, 6]),
32
+ @table.sort_indices("-number1"))
33
+ end
34
+
35
+ test("Symbol, -String") do
36
+ assert_equal(Arrow::UInt64Array.new([6, 5, 4, 1, 2, 0, 3]),
37
+ @table.sort_indices([:number1, "-number2"]))
38
+ end
39
+ end
40
+ end
@@ -0,0 +1,81 @@
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 SortKeyTest < Test::Unit::TestCase
19
+ sub_test_case(".resolve") do
20
+ test("SortKey") do
21
+ assert_equal(Arrow::SortKey.new("-count"),
22
+ Arrow::SortKey.resolve(Arrow::SortKey.new("-count")))
23
+ end
24
+
25
+ test("-String") do
26
+ assert_equal(Arrow::SortKey.new("-count"),
27
+ Arrow::SortKey.resolve("-count"))
28
+ end
29
+
30
+ test("Symbol, Symbol") do
31
+ assert_equal(Arrow::SortKey.new("-count"),
32
+ Arrow::SortKey.resolve(:count, :desc))
33
+ end
34
+ end
35
+
36
+ sub_test_case("#initialize") do
37
+ test("String") do
38
+ assert_equal("+count",
39
+ Arrow::SortKey.new("count").to_s)
40
+ end
41
+
42
+ test("+String") do
43
+ assert_equal("+count",
44
+ Arrow::SortKey.new("+count").to_s)
45
+ end
46
+
47
+ test("-String") do
48
+ assert_equal("-count",
49
+ Arrow::SortKey.new("-count").to_s)
50
+ end
51
+
52
+ test("Symbol") do
53
+ assert_equal("+-count",
54
+ Arrow::SortKey.new(:"-count").to_s)
55
+ end
56
+
57
+ test("String, Symbol") do
58
+ assert_equal("--count",
59
+ Arrow::SortKey.new("-count", :desc).to_s)
60
+ end
61
+
62
+ test("String, String") do
63
+ assert_equal("--count",
64
+ Arrow::SortKey.new("-count", "desc").to_s)
65
+ end
66
+
67
+ test("String, SortOrder") do
68
+ assert_equal("--count",
69
+ Arrow::SortKey.new("-count",
70
+ Arrow::SortOrder::DESCENDING).to_s)
71
+ end
72
+ end
73
+
74
+ sub_test_case("#to_s") do
75
+ test("recreatable") do
76
+ key = Arrow::SortKey.new("-count", :desc)
77
+ assert_equal(key,
78
+ Arrow::SortKey.new(key.to_s))
79
+ end
80
+ end
81
+ end
@@ -0,0 +1,58 @@
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 SortOptionsTest < Test::Unit::TestCase
19
+ sub_test_case("#initialize") do
20
+ test("none") do
21
+ options = Arrow::SortOptions.new
22
+ assert_equal([],
23
+ options.sort_keys.collect(&:to_s))
24
+ end
25
+
26
+ test("-String, Symbol") do
27
+ options = Arrow::SortOptions.new("-count", :age)
28
+ assert_equal(["-count", "+age"],
29
+ options.sort_keys.collect(&:to_s))
30
+ end
31
+ end
32
+
33
+ sub_test_case("instance methods") do
34
+ setup do
35
+ @options = Arrow::SortOptions.new
36
+ end
37
+
38
+ sub_test_case("#add_sort_key") do
39
+ test("-String") do
40
+ @options.add_sort_key("-count")
41
+ assert_equal(["-count"],
42
+ @options.sort_keys.collect(&:to_s))
43
+ end
44
+
45
+ test("-String, Symbol") do
46
+ @options.add_sort_key("-count", :desc)
47
+ assert_equal(["--count"],
48
+ @options.sort_keys.collect(&:to_s))
49
+ end
50
+
51
+ test("SortKey") do
52
+ @options.add_sort_key(Arrow::SortKey.new("-count"))
53
+ assert_equal(["-count"],
54
+ @options.sort_keys.collect(&:to_s))
55
+ end
56
+ end
57
+ end
58
+ end
@@ -28,12 +28,12 @@ class SparseUnionDataTypeTest < Test::Unit::TestCase
28
28
  end
29
29
 
30
30
  test("ordered arguments") do
31
- assert_equal("union[sparse]<visible: bool=2, count: int32=9>",
31
+ assert_equal("sparse_union<visible: bool=2, count: int32=9>",
32
32
  Arrow::SparseUnionDataType.new(@fields, [2, 9]).to_s)
33
33
  end
34
34
 
35
35
  test("description") do
36
- assert_equal("union[sparse]<visible: bool=2, count: int32=9>",
36
+ assert_equal("sparse_union<visible: bool=2, count: int32=9>",
37
37
  Arrow::SparseUnionDataType.new(fields: @fields,
38
38
  type_codes: [2, 9]).to_s)
39
39
  end