red-arrow 4.0.1 → 5.0.0
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.
- checksums.yaml +4 -4
- data/ext/arrow/arrow.cpp +3 -0
- data/ext/arrow/memory-view.cpp +311 -0
- data/ext/arrow/memory-view.hpp +26 -0
- data/lib/arrow/array.rb +12 -0
- data/lib/arrow/buffer.rb +10 -6
- data/lib/arrow/constructor-arguments-gc-guardable.rb +25 -0
- data/lib/arrow/datum.rb +98 -0
- data/lib/arrow/equal-options.rb +38 -0
- data/lib/arrow/loader.rb +31 -0
- data/lib/arrow/scalar.rb +32 -0
- data/lib/arrow/table.rb +2 -2
- data/lib/arrow/version.rb +1 -1
- data/red-arrow.gemspec +2 -1
- data/test/helper.rb +1 -0
- data/test/test-array.rb +34 -0
- data/test/test-boolean-scalar.rb +26 -0
- data/test/test-float-scalar.rb +46 -0
- data/test/test-function.rb +176 -0
- data/test/test-memory-view.rb +434 -0
- metadata +101 -73
@@ -0,0 +1,38 @@
|
|
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
|
+
module Arrow
|
19
|
+
class EqualOptions
|
20
|
+
class << self
|
21
|
+
# @api private
|
22
|
+
def try_convert(value)
|
23
|
+
case value
|
24
|
+
when Hash
|
25
|
+
options = new
|
26
|
+
value.each do |k, v|
|
27
|
+
setter = :"#{k}="
|
28
|
+
return unless options.respond_to?(setter)
|
29
|
+
options.__send__(setter, v)
|
30
|
+
end
|
31
|
+
options
|
32
|
+
else
|
33
|
+
nil
|
34
|
+
end
|
35
|
+
end
|
36
|
+
end
|
37
|
+
end
|
38
|
+
end
|
data/lib/arrow/loader.rb
CHANGED
@@ -29,6 +29,7 @@ module Arrow
|
|
29
29
|
def post_load(repository, namespace)
|
30
30
|
require_libraries
|
31
31
|
require_extension_library
|
32
|
+
gc_guard
|
32
33
|
end
|
33
34
|
|
34
35
|
def require_libraries
|
@@ -52,6 +53,7 @@ module Arrow
|
|
52
53
|
require "arrow/date32-array-builder"
|
53
54
|
require "arrow/date64-array"
|
54
55
|
require "arrow/date64-array-builder"
|
56
|
+
require "arrow/datum"
|
55
57
|
require "arrow/decimal128"
|
56
58
|
require "arrow/decimal128-array"
|
57
59
|
require "arrow/decimal128-array-builder"
|
@@ -63,6 +65,7 @@ module Arrow
|
|
63
65
|
require "arrow/dense-union-data-type"
|
64
66
|
require "arrow/dictionary-array"
|
65
67
|
require "arrow/dictionary-data-type"
|
68
|
+
require "arrow/equal-options"
|
66
69
|
require "arrow/field"
|
67
70
|
require "arrow/file-output-stream"
|
68
71
|
require "arrow/fixed-size-binary-array"
|
@@ -80,6 +83,7 @@ module Arrow
|
|
80
83
|
require "arrow/record-batch-iterator"
|
81
84
|
require "arrow/record-batch-stream-reader"
|
82
85
|
require "arrow/rolling-window"
|
86
|
+
require "arrow/scalar"
|
83
87
|
require "arrow/schema"
|
84
88
|
require "arrow/slicer"
|
85
89
|
require "arrow/sort-key"
|
@@ -112,6 +116,27 @@ module Arrow
|
|
112
116
|
require "arrow.so"
|
113
117
|
end
|
114
118
|
|
119
|
+
def gc_guard
|
120
|
+
require "arrow/constructor-arguments-gc-guardable"
|
121
|
+
|
122
|
+
[
|
123
|
+
@base_module::BinaryScalar,
|
124
|
+
@base_module::Buffer,
|
125
|
+
@base_module::DenseUnionScalar,
|
126
|
+
@base_module::FixedSizeBinaryScalar,
|
127
|
+
@base_module::LargeBinaryScalar,
|
128
|
+
@base_module::LargeListScalar,
|
129
|
+
@base_module::LargeStringScalar,
|
130
|
+
@base_module::ListScalar,
|
131
|
+
@base_module::MapScalar,
|
132
|
+
@base_module::SparseUnionScalar,
|
133
|
+
@base_module::StringScalar,
|
134
|
+
@base_module::StructScalar,
|
135
|
+
].each do |klass|
|
136
|
+
klass.prepend(ConstructorArgumentsGCGuardable)
|
137
|
+
end
|
138
|
+
end
|
139
|
+
|
115
140
|
def load_object_info(info)
|
116
141
|
super
|
117
142
|
|
@@ -164,6 +189,12 @@ module Arrow
|
|
164
189
|
method_name = "dup"
|
165
190
|
end
|
166
191
|
super(info, klass, method_name)
|
192
|
+
when "Arrow::BooleanScalar"
|
193
|
+
case method_name
|
194
|
+
when "value?"
|
195
|
+
method_name = "value"
|
196
|
+
end
|
197
|
+
super(info, klass, method_name)
|
167
198
|
else
|
168
199
|
super
|
169
200
|
end
|
data/lib/arrow/scalar.rb
ADDED
@@ -0,0 +1,32 @@
|
|
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
|
+
module Arrow
|
19
|
+
class Scalar
|
20
|
+
# @param other [Arrow::Scalar] The scalar to be compared.
|
21
|
+
# @param options [Arrow::EqualOptions, Hash] (nil)
|
22
|
+
# The options to custom how to compare.
|
23
|
+
#
|
24
|
+
# @return [Boolean]
|
25
|
+
# `true` if both of them have the same data, `false` otherwise.
|
26
|
+
#
|
27
|
+
# @since 5.0.0
|
28
|
+
def equal_scalar?(other, options=nil)
|
29
|
+
equal_options(other, options)
|
30
|
+
end
|
31
|
+
end
|
32
|
+
end
|
data/lib/arrow/table.rb
CHANGED
data/lib/arrow/version.rb
CHANGED
data/red-arrow.gemspec
CHANGED
@@ -48,13 +48,14 @@ Gem::Specification.new do |spec|
|
|
48
48
|
|
49
49
|
spec.add_runtime_dependency("bigdecimal", ">= 2.0.3")
|
50
50
|
spec.add_runtime_dependency("extpp", ">= 0.0.7")
|
51
|
-
spec.add_runtime_dependency("gio2", ">= 3.
|
51
|
+
spec.add_runtime_dependency("gio2", ">= 3.4.5")
|
52
52
|
spec.add_runtime_dependency("native-package-installer")
|
53
53
|
spec.add_runtime_dependency("pkg-config")
|
54
54
|
|
55
55
|
spec.add_development_dependency("benchmark-driver")
|
56
56
|
spec.add_development_dependency("bundler")
|
57
57
|
spec.add_development_dependency("faker")
|
58
|
+
spec.add_development_dependency("fiddle", ">= 1.0.9")
|
58
59
|
spec.add_development_dependency("rake")
|
59
60
|
spec.add_development_dependency("redcarpet")
|
60
61
|
spec.add_development_dependency("test-unit")
|
data/test/helper.rb
CHANGED
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]),
|
@@ -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
|
@@ -0,0 +1,46 @@
|
|
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 FloatScalarTest < Test::Unit::TestCase
|
19
|
+
sub_test_case("#equal_scalar?") do
|
20
|
+
test("no options") do
|
21
|
+
scalar1 = Arrow::FloatScalar.new(1.1)
|
22
|
+
scalar2 = Arrow::FloatScalar.new(1.1000001)
|
23
|
+
assert do
|
24
|
+
not scalar1.equal_scalar?(scalar2)
|
25
|
+
end
|
26
|
+
end
|
27
|
+
|
28
|
+
test(":approx") do
|
29
|
+
scalar1 = Arrow::FloatScalar.new(1.1)
|
30
|
+
scalar2 = Arrow::FloatScalar.new(1.1000001)
|
31
|
+
assert do
|
32
|
+
scalar1.equal_scalar?(scalar2, approx: true)
|
33
|
+
end
|
34
|
+
end
|
35
|
+
|
36
|
+
test(":absolute_tolerance") do
|
37
|
+
scalar1 = Arrow::FloatScalar.new(1.1)
|
38
|
+
scalar2 = Arrow::FloatScalar.new(1.1001)
|
39
|
+
assert do
|
40
|
+
scalar1.equal_scalar?(scalar2,
|
41
|
+
approx: true,
|
42
|
+
absolute_tolerance: 0.001)
|
43
|
+
end
|
44
|
+
end
|
45
|
+
end
|
46
|
+
end
|
@@ -0,0 +1,176 @@
|
|
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 FunctionTest < Test::Unit::TestCase
|
19
|
+
sub_test_case("#execute") do
|
20
|
+
test("Arrow::Array") do
|
21
|
+
or_function = Arrow::Function.find("or")
|
22
|
+
args = [
|
23
|
+
Arrow::BooleanArray.new([true, false, false]),
|
24
|
+
Arrow::BooleanArray.new([true, false, true]),
|
25
|
+
]
|
26
|
+
assert_equal([true, false, true],
|
27
|
+
or_function.execute(args).value.to_a)
|
28
|
+
end
|
29
|
+
|
30
|
+
test("Array") do
|
31
|
+
or_function = Arrow::Function.find("or")
|
32
|
+
args = [
|
33
|
+
[true, false, false],
|
34
|
+
[true, false, true],
|
35
|
+
]
|
36
|
+
assert_equal([true, false, true],
|
37
|
+
or_function.execute(args).value.to_a)
|
38
|
+
end
|
39
|
+
|
40
|
+
test("Arrow::ChunkedArray") do
|
41
|
+
or_function = Arrow::Function.find("or")
|
42
|
+
args = [
|
43
|
+
Arrow::ChunkedArray.new([
|
44
|
+
Arrow::BooleanArray.new([true]),
|
45
|
+
Arrow::BooleanArray.new([false, false]),
|
46
|
+
]),
|
47
|
+
Arrow::ChunkedArray.new([
|
48
|
+
Arrow::BooleanArray.new([true, false]),
|
49
|
+
Arrow::BooleanArray.new([true]),
|
50
|
+
]),
|
51
|
+
]
|
52
|
+
assert_equal([true, false, true],
|
53
|
+
or_function.execute(args).value.to_a)
|
54
|
+
end
|
55
|
+
|
56
|
+
test("Arrow::Scalar") do
|
57
|
+
add_function = Arrow::Function.find("add")
|
58
|
+
args = [
|
59
|
+
Arrow::Int8Array.new([1, 2, 3]),
|
60
|
+
Arrow::Int8Scalar.new(5),
|
61
|
+
]
|
62
|
+
assert_equal([6, 7, 8],
|
63
|
+
add_function.execute(args).value.to_a)
|
64
|
+
end
|
65
|
+
|
66
|
+
test("Integer") do
|
67
|
+
add_function = Arrow::Function.find("add")
|
68
|
+
args = [
|
69
|
+
[1, 2, 3],
|
70
|
+
5,
|
71
|
+
]
|
72
|
+
assert_equal([6, 7, 8],
|
73
|
+
add_function.execute(args).value.to_a)
|
74
|
+
end
|
75
|
+
|
76
|
+
test("Float") do
|
77
|
+
add_function = Arrow::Function.find("add")
|
78
|
+
args = [
|
79
|
+
[1, 2, 3],
|
80
|
+
5.1,
|
81
|
+
]
|
82
|
+
assert_equal([6.1, 7.1, 8.1],
|
83
|
+
add_function.execute(args).value.to_a)
|
84
|
+
end
|
85
|
+
|
86
|
+
test("true") do
|
87
|
+
and_function = Arrow::Function.find("and")
|
88
|
+
args = [
|
89
|
+
Arrow::BooleanArray.new([true, false, false]),
|
90
|
+
true,
|
91
|
+
]
|
92
|
+
assert_equal([true, false, false],
|
93
|
+
and_function.execute(args).value.to_a)
|
94
|
+
end
|
95
|
+
|
96
|
+
test("false") do
|
97
|
+
or_function = Arrow::Function.find("or")
|
98
|
+
args = [
|
99
|
+
Arrow::BooleanArray.new([true, false, false]),
|
100
|
+
false,
|
101
|
+
]
|
102
|
+
assert_equal([true, false, false],
|
103
|
+
or_function.execute(args).value.to_a)
|
104
|
+
end
|
105
|
+
|
106
|
+
test("String") do
|
107
|
+
ascii_upper_function = Arrow::Function.find("ascii_upper")
|
108
|
+
args = [
|
109
|
+
"Hello",
|
110
|
+
]
|
111
|
+
assert_equal("HELLO",
|
112
|
+
ascii_upper_function.execute(args).value.to_s)
|
113
|
+
end
|
114
|
+
|
115
|
+
test("Date") do
|
116
|
+
cast_function = Arrow::Function.find("cast")
|
117
|
+
date = Date.new(2021, 6, 12)
|
118
|
+
args = [date]
|
119
|
+
options = Arrow::CastOptions.new
|
120
|
+
options.to_data_type = Arrow::TimestampDataType.new(:second)
|
121
|
+
time = Time.utc(date.year,
|
122
|
+
date.month,
|
123
|
+
date.day)
|
124
|
+
assert_equal(Arrow::TimestampScalar.new(options.to_data_type,
|
125
|
+
time.to_i),
|
126
|
+
cast_function.execute(args, options).value)
|
127
|
+
end
|
128
|
+
|
129
|
+
test("Arrow::Time: second") do
|
130
|
+
cast_function = Arrow::Function.find("cast")
|
131
|
+
arrow_time = Arrow::Time.new(Arrow::TimeUnit::SECOND,
|
132
|
+
# 00:10:00
|
133
|
+
60 * 10)
|
134
|
+
args = [arrow_time]
|
135
|
+
options = Arrow::CastOptions.new
|
136
|
+
options.to_data_type = Arrow::Time64DataType.new(:micro)
|
137
|
+
assert_equal(Arrow::Time64Scalar.new(options.to_data_type,
|
138
|
+
# 00:10:00.000000
|
139
|
+
60 * 10 * 1000 * 1000),
|
140
|
+
cast_function.execute(args, options).value)
|
141
|
+
end
|
142
|
+
|
143
|
+
test("Arrow::Time: micro") do
|
144
|
+
cast_function = Arrow::Function.find("cast")
|
145
|
+
arrow_time = Arrow::Time.new(Arrow::TimeUnit::MICRO,
|
146
|
+
# 00:10:00.000000
|
147
|
+
60 * 10 * 1000 * 1000)
|
148
|
+
args = [arrow_time]
|
149
|
+
options = Arrow::CastOptions.new
|
150
|
+
options.to_data_type = Arrow::Time32DataType.new(:second)
|
151
|
+
options.allow_time_truncate = true
|
152
|
+
assert_equal(Arrow::Time32Scalar.new(options.to_data_type,
|
153
|
+
# 00:10:00
|
154
|
+
60 * 10),
|
155
|
+
cast_function.execute(args, options).value)
|
156
|
+
end
|
157
|
+
|
158
|
+
test("Time") do
|
159
|
+
cast_function = Arrow::Function.find("cast")
|
160
|
+
time = Time.utc(2021, 6, 12, 1, 2, 3, 1)
|
161
|
+
args = [time]
|
162
|
+
options = Arrow::CastOptions.new
|
163
|
+
options.to_data_type = Arrow::TimestampDataType.new(:second)
|
164
|
+
options.allow_time_truncate = true
|
165
|
+
time = Time.utc(time.year,
|
166
|
+
time.month,
|
167
|
+
time.day,
|
168
|
+
time.hour,
|
169
|
+
time.min,
|
170
|
+
time.sec)
|
171
|
+
assert_equal(Arrow::TimestampScalar.new(options.to_data_type,
|
172
|
+
time.to_i),
|
173
|
+
cast_function.execute(args, options).value)
|
174
|
+
end
|
175
|
+
end
|
176
|
+
end
|