red-arrow 0.4.1 → 0.8.0
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/README.md +3 -2
- data/dependency-check/Rakefile +1 -1
- data/doc/text/news.md +37 -1
- data/image/red-arrow.png +0 -0
- data/lib/arrow/array-builder.rb +38 -34
- data/lib/arrow/array.rb +11 -5
- data/lib/arrow/chunked-array.rb +18 -2
- data/lib/arrow/column.rb +4 -0
- data/lib/arrow/csv-reader.rb +162 -0
- data/lib/arrow/loader.rb +14 -12
- data/lib/arrow/slicer.rb +391 -0
- data/lib/arrow/table-formatter.rb +88 -0
- data/lib/arrow/table.rb +309 -3
- data/lib/arrow/timestamp-array.rb +69 -0
- data/lib/arrow/version.rb +1 -1
- data/red-arrow.gemspec +1 -0
- data/test/fixture/with-header.csv +4 -0
- data/test/fixture/without-header.csv +3 -0
- data/test/helper.rb +3 -0
- data/test/helper/fixture.rb +25 -0
- data/test/test-array.rb +6 -0
- data/test/test-column.rb +2 -2
- data/test/test-csv-reader.rb +90 -0
- data/test/test-slicer.rb +401 -0
- data/test/test-table.rb +321 -13
- data/test/test-timestamp-array.rb +23 -0
- metadata +24 -7
data/test/test-array.rb
CHANGED
@@ -26,4 +26,10 @@ class ArrayTest < Test::Unit::TestCase
|
|
26
26
|
assert_equal([true, false, nil, true],
|
27
27
|
array.to_a)
|
28
28
|
end
|
29
|
+
|
30
|
+
test("#[]") do
|
31
|
+
array = Arrow::BooleanArray.new([true, false, nil, true])
|
32
|
+
assert_equal([true, false, nil, true],
|
33
|
+
[array[0], array[1], array[2], array[3]])
|
34
|
+
end
|
29
35
|
end
|
data/test/test-column.rb
CHANGED
@@ -14,11 +14,11 @@
|
|
14
14
|
|
15
15
|
class ColumnTest < Test::Unit::TestCase
|
16
16
|
test("#each") do
|
17
|
-
|
17
|
+
arrays = [
|
18
18
|
Arrow::BooleanArray.new([true, false]),
|
19
19
|
Arrow::BooleanArray.new([nil, true]),
|
20
20
|
]
|
21
|
-
chunked_array = Arrow::ChunkedArray.new(
|
21
|
+
chunked_array = Arrow::ChunkedArray.new(arrays)
|
22
22
|
column = Arrow::Column.new(Arrow::Field.new("visible", :boolean),
|
23
23
|
chunked_array)
|
24
24
|
assert_equal([true, false, nil, true],
|
@@ -0,0 +1,90 @@
|
|
1
|
+
# Copyright 2017-2018 Kouhei Sutou <kou@clear-code.com>
|
2
|
+
#
|
3
|
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
4
|
+
# you may not use this file except in compliance with the License.
|
5
|
+
# You may obtain a copy of the License at
|
6
|
+
#
|
7
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
8
|
+
#
|
9
|
+
# Unless required by applicable law or agreed to in writing, software
|
10
|
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
11
|
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
12
|
+
# See the License for the specific language governing permissions and
|
13
|
+
# limitations under the License.
|
14
|
+
|
15
|
+
class CSVReaderTest < Test::Unit::TestCase
|
16
|
+
include Helper::Fixture
|
17
|
+
|
18
|
+
sub_test_case(".read") do
|
19
|
+
test("String: data: with header") do
|
20
|
+
data = fixture_path("with-header.csv").read
|
21
|
+
assert_equal(<<-TABLE, Arrow::CSVReader.read(data).to_s)
|
22
|
+
name score
|
23
|
+
0 alice 10
|
24
|
+
1 bob 29
|
25
|
+
2 chris -1
|
26
|
+
TABLE
|
27
|
+
end
|
28
|
+
|
29
|
+
test("String: data: without header") do
|
30
|
+
data = fixture_path("without-header.csv").read
|
31
|
+
assert_equal(<<-TABLE, Arrow::CSVReader.read(data).to_s)
|
32
|
+
0 1
|
33
|
+
0 alice 10
|
34
|
+
1 bob 29
|
35
|
+
2 chris -1
|
36
|
+
TABLE
|
37
|
+
end
|
38
|
+
|
39
|
+
test("String: path: with header") do
|
40
|
+
path = fixture_path("with-header.csv").to_s
|
41
|
+
assert_equal(<<-TABLE, Arrow::CSVReader.read(path).to_s)
|
42
|
+
name score
|
43
|
+
0 alice 10
|
44
|
+
1 bob 29
|
45
|
+
2 chris -1
|
46
|
+
TABLE
|
47
|
+
end
|
48
|
+
|
49
|
+
test("String: path: without header") do
|
50
|
+
path = fixture_path("without-header.csv").to_s
|
51
|
+
assert_equal(<<-TABLE, Arrow::CSVReader.read(path).to_s)
|
52
|
+
0 1
|
53
|
+
0 alice 10
|
54
|
+
1 bob 29
|
55
|
+
2 chris -1
|
56
|
+
TABLE
|
57
|
+
end
|
58
|
+
|
59
|
+
test("Pathname: with header") do
|
60
|
+
path = fixture_path("with-header.csv")
|
61
|
+
assert_equal(<<-TABLE, Arrow::CSVReader.read(path).to_s)
|
62
|
+
name score
|
63
|
+
0 alice 10
|
64
|
+
1 bob 29
|
65
|
+
2 chris -1
|
66
|
+
TABLE
|
67
|
+
end
|
68
|
+
|
69
|
+
test("Pathname: without header") do
|
70
|
+
path = fixture_path("without-header.csv")
|
71
|
+
assert_equal(<<-TABLE, Arrow::CSVReader.read(path).to_s)
|
72
|
+
0 1
|
73
|
+
0 alice 10
|
74
|
+
1 bob 29
|
75
|
+
2 chris -1
|
76
|
+
TABLE
|
77
|
+
end
|
78
|
+
|
79
|
+
test("CSV") do
|
80
|
+
CSV.open(fixture_path("with-header.csv").to_s, headers: true) do |csv|
|
81
|
+
assert_equal(<<-TABLE, Arrow::CSVReader.read(csv).to_s)
|
82
|
+
name score
|
83
|
+
0 alice 10
|
84
|
+
1 bob 29
|
85
|
+
2 chris -1
|
86
|
+
TABLE
|
87
|
+
end
|
88
|
+
end
|
89
|
+
end
|
90
|
+
end
|
data/test/test-slicer.rb
ADDED
@@ -0,0 +1,401 @@
|
|
1
|
+
# Copyright 2017-2018 Kouhei Sutou <kou@clear-code.com>
|
2
|
+
#
|
3
|
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
4
|
+
# you may not use this file except in compliance with the License.
|
5
|
+
# You may obtain a copy of the License at
|
6
|
+
#
|
7
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
8
|
+
#
|
9
|
+
# Unless required by applicable law or agreed to in writing, software
|
10
|
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
11
|
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
12
|
+
# See the License for the specific language governing permissions and
|
13
|
+
# limitations under the License.
|
14
|
+
|
15
|
+
class SlicerTest < Test::Unit::TestCase
|
16
|
+
def setup
|
17
|
+
@count_field = Arrow::Field.new("count", :uint32)
|
18
|
+
@visible_field = Arrow::Field.new("visible", :boolean)
|
19
|
+
schema = Arrow::Schema.new([@count_field, @visible_field])
|
20
|
+
count_arrays = [
|
21
|
+
Arrow::UInt32Array.new([0, 1, 2]),
|
22
|
+
Arrow::UInt32Array.new([4, 8, 16]),
|
23
|
+
Arrow::UInt32Array.new([32, 64, nil]),
|
24
|
+
Arrow::UInt32Array.new([256]),
|
25
|
+
]
|
26
|
+
visible_arrays = [
|
27
|
+
Arrow::BooleanArray.new([nil, true, false, nil]),
|
28
|
+
Arrow::BooleanArray.new([true]),
|
29
|
+
Arrow::BooleanArray.new([true, false]),
|
30
|
+
Arrow::BooleanArray.new([nil]),
|
31
|
+
Arrow::BooleanArray.new([nil]),
|
32
|
+
Arrow::BooleanArray.new([true]),
|
33
|
+
]
|
34
|
+
@count_array = Arrow::ChunkedArray.new(count_arrays)
|
35
|
+
@visible_array = Arrow::ChunkedArray.new(visible_arrays)
|
36
|
+
@count_column = Arrow::Column.new(@count_field, @count_array)
|
37
|
+
@visible_column = Arrow::Column.new(@visible_field, @visible_array)
|
38
|
+
@table = Arrow::Table.new(schema, [@count_column, @visible_column])
|
39
|
+
end
|
40
|
+
|
41
|
+
sub_test_case("column") do
|
42
|
+
test("BooleanArray") do
|
43
|
+
sliced_table = @table.slice do |slicer|
|
44
|
+
slicer.visible
|
45
|
+
end
|
46
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
47
|
+
count visible
|
48
|
+
0 1 true
|
49
|
+
1 8 true
|
50
|
+
2 16 true
|
51
|
+
3 256 true
|
52
|
+
TABLE
|
53
|
+
end
|
54
|
+
|
55
|
+
test("not BooleanArray") do
|
56
|
+
sliced_table = @table.slice do |slicer|
|
57
|
+
slicer.count
|
58
|
+
end
|
59
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
60
|
+
count visible
|
61
|
+
0 1 true
|
62
|
+
1 2 false
|
63
|
+
2 4
|
64
|
+
3 8 true
|
65
|
+
4 16 true
|
66
|
+
5 32 false
|
67
|
+
6 64
|
68
|
+
7 256 true
|
69
|
+
TABLE
|
70
|
+
end
|
71
|
+
end
|
72
|
+
|
73
|
+
sub_test_case("!column") do
|
74
|
+
test("BooleanArray") do
|
75
|
+
sliced_table = @table.slice do |slicer|
|
76
|
+
!slicer.visible
|
77
|
+
end
|
78
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
79
|
+
count visible
|
80
|
+
0 2 false
|
81
|
+
1 32 false
|
82
|
+
TABLE
|
83
|
+
end
|
84
|
+
|
85
|
+
test("not BooleanArray") do
|
86
|
+
sliced_table = @table.slice do |slicer|
|
87
|
+
!slicer.count
|
88
|
+
end
|
89
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
90
|
+
count visible
|
91
|
+
0 0
|
92
|
+
TABLE
|
93
|
+
end
|
94
|
+
end
|
95
|
+
|
96
|
+
test("column.null?") do
|
97
|
+
sliced_table = @table.slice do |slicer|
|
98
|
+
slicer.visible.null?
|
99
|
+
end
|
100
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
101
|
+
count visible
|
102
|
+
0 0
|
103
|
+
1 4
|
104
|
+
2 64
|
105
|
+
3
|
106
|
+
TABLE
|
107
|
+
end
|
108
|
+
|
109
|
+
sub_test_case("column ==") do
|
110
|
+
test("nil") do
|
111
|
+
sliced_table = @table.slice do |slicer|
|
112
|
+
slicer.visible == nil
|
113
|
+
end
|
114
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
115
|
+
count visible
|
116
|
+
0 0
|
117
|
+
1 4
|
118
|
+
2 64
|
119
|
+
3
|
120
|
+
TABLE
|
121
|
+
end
|
122
|
+
|
123
|
+
test("value") do
|
124
|
+
sliced_table = @table.slice do |slicer|
|
125
|
+
slicer.visible == true
|
126
|
+
end
|
127
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
128
|
+
count visible
|
129
|
+
0 1 true
|
130
|
+
1 8 true
|
131
|
+
2 16 true
|
132
|
+
3 256 true
|
133
|
+
TABLE
|
134
|
+
end
|
135
|
+
end
|
136
|
+
|
137
|
+
sub_test_case("!(column ==)") do
|
138
|
+
test("nil") do
|
139
|
+
sliced_table = @table.slice do |slicer|
|
140
|
+
!(slicer.visible == nil)
|
141
|
+
end
|
142
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
143
|
+
count visible
|
144
|
+
0 1 true
|
145
|
+
1 2 false
|
146
|
+
2 8 true
|
147
|
+
3 16 true
|
148
|
+
4 32 false
|
149
|
+
5 256 true
|
150
|
+
TABLE
|
151
|
+
end
|
152
|
+
|
153
|
+
test("value") do
|
154
|
+
sliced_table = @table.slice do |slicer|
|
155
|
+
!(slicer.visible == true)
|
156
|
+
end
|
157
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
158
|
+
count visible
|
159
|
+
0 2 false
|
160
|
+
1 32 false
|
161
|
+
TABLE
|
162
|
+
end
|
163
|
+
end
|
164
|
+
|
165
|
+
sub_test_case("column !=") do
|
166
|
+
test("nil") do
|
167
|
+
sliced_table = @table.slice do |slicer|
|
168
|
+
slicer.visible != nil
|
169
|
+
end
|
170
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
171
|
+
count visible
|
172
|
+
0 1 true
|
173
|
+
1 2 false
|
174
|
+
2 8 true
|
175
|
+
3 16 true
|
176
|
+
4 32 false
|
177
|
+
5 256 true
|
178
|
+
TABLE
|
179
|
+
end
|
180
|
+
|
181
|
+
test("value") do
|
182
|
+
sliced_table = @table.slice do |slicer|
|
183
|
+
slicer.visible != true
|
184
|
+
end
|
185
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
186
|
+
count visible
|
187
|
+
0 2 false
|
188
|
+
1 32 false
|
189
|
+
TABLE
|
190
|
+
end
|
191
|
+
end
|
192
|
+
|
193
|
+
test("column < value") do
|
194
|
+
sliced_table = @table.slice do |slicer|
|
195
|
+
slicer.count < 16
|
196
|
+
end
|
197
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
198
|
+
count visible
|
199
|
+
0 0
|
200
|
+
1 1 true
|
201
|
+
2 2 false
|
202
|
+
3 4
|
203
|
+
4 8 true
|
204
|
+
TABLE
|
205
|
+
end
|
206
|
+
|
207
|
+
test("!(column < value)") do
|
208
|
+
sliced_table = @table.slice do |slicer|
|
209
|
+
!(slicer.count < 16)
|
210
|
+
end
|
211
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
212
|
+
count visible
|
213
|
+
0 16 true
|
214
|
+
1 32 false
|
215
|
+
2 64
|
216
|
+
3 256 true
|
217
|
+
TABLE
|
218
|
+
end
|
219
|
+
|
220
|
+
test("column <= value") do
|
221
|
+
sliced_table = @table.slice do |slicer|
|
222
|
+
slicer.count <= 16
|
223
|
+
end
|
224
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
225
|
+
count visible
|
226
|
+
0 0
|
227
|
+
1 1 true
|
228
|
+
2 2 false
|
229
|
+
3 4
|
230
|
+
4 8 true
|
231
|
+
5 16 true
|
232
|
+
TABLE
|
233
|
+
end
|
234
|
+
|
235
|
+
test("!(column <= value)") do
|
236
|
+
sliced_table = @table.slice do |slicer|
|
237
|
+
!(slicer.count <= 16)
|
238
|
+
end
|
239
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
240
|
+
count visible
|
241
|
+
0 32 false
|
242
|
+
1 64
|
243
|
+
2 256 true
|
244
|
+
TABLE
|
245
|
+
end
|
246
|
+
|
247
|
+
test("column > value") do
|
248
|
+
sliced_table = @table.slice do |slicer|
|
249
|
+
slicer.count > 16
|
250
|
+
end
|
251
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
252
|
+
count visible
|
253
|
+
0 32 false
|
254
|
+
1 64
|
255
|
+
2 256 true
|
256
|
+
TABLE
|
257
|
+
end
|
258
|
+
|
259
|
+
test("!(column > value)") do
|
260
|
+
sliced_table = @table.slice do |slicer|
|
261
|
+
!(slicer.count > 16)
|
262
|
+
end
|
263
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
264
|
+
count visible
|
265
|
+
0 0
|
266
|
+
1 1 true
|
267
|
+
2 2 false
|
268
|
+
3 4
|
269
|
+
4 8 true
|
270
|
+
5 16 true
|
271
|
+
TABLE
|
272
|
+
end
|
273
|
+
|
274
|
+
test("column >= value") do
|
275
|
+
sliced_table = @table.slice do |slicer|
|
276
|
+
slicer.count >= 16
|
277
|
+
end
|
278
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
279
|
+
count visible
|
280
|
+
0 16 true
|
281
|
+
1 32 false
|
282
|
+
2 64
|
283
|
+
3 256 true
|
284
|
+
TABLE
|
285
|
+
end
|
286
|
+
|
287
|
+
test("!(column >= value)") do
|
288
|
+
sliced_table = @table.slice do |slicer|
|
289
|
+
!(slicer.count >= 16)
|
290
|
+
end
|
291
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
292
|
+
count visible
|
293
|
+
0 0
|
294
|
+
1 1 true
|
295
|
+
2 2 false
|
296
|
+
3 4
|
297
|
+
4 8 true
|
298
|
+
TABLE
|
299
|
+
end
|
300
|
+
|
301
|
+
test("condition & condition") do
|
302
|
+
sliced_table = @table.slice do |slicer|
|
303
|
+
slicer.visible & (slicer.count >= 16)
|
304
|
+
end
|
305
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
306
|
+
count visible
|
307
|
+
0 16 true
|
308
|
+
1 256 true
|
309
|
+
TABLE
|
310
|
+
end
|
311
|
+
|
312
|
+
test("condition | condition") do
|
313
|
+
sliced_table = @table.slice do |slicer|
|
314
|
+
slicer.visible | (slicer.count >= 16)
|
315
|
+
end
|
316
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
317
|
+
count visible
|
318
|
+
0 1 true
|
319
|
+
1 8 true
|
320
|
+
2 16 true
|
321
|
+
3 32 false
|
322
|
+
4 256 true
|
323
|
+
TABLE
|
324
|
+
end
|
325
|
+
|
326
|
+
test("condition ^ condition") do
|
327
|
+
sliced_table = @table.slice do |slicer|
|
328
|
+
slicer.visible ^ (slicer.count >= 16)
|
329
|
+
end
|
330
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
331
|
+
count visible
|
332
|
+
0 1 true
|
333
|
+
1 8 true
|
334
|
+
2 32 false
|
335
|
+
TABLE
|
336
|
+
end
|
337
|
+
|
338
|
+
test("select") do
|
339
|
+
sliced_table = @table.slice do |slicer|
|
340
|
+
slicer.visible.select do |value|
|
341
|
+
value.nil? or value
|
342
|
+
end
|
343
|
+
end
|
344
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
345
|
+
count visible
|
346
|
+
0 0
|
347
|
+
1 1 true
|
348
|
+
2 4
|
349
|
+
3 8 true
|
350
|
+
4 16 true
|
351
|
+
5 64
|
352
|
+
6
|
353
|
+
7 256 true
|
354
|
+
TABLE
|
355
|
+
end
|
356
|
+
|
357
|
+
test("!select") do
|
358
|
+
sliced_table = @table.slice do |slicer|
|
359
|
+
!slicer.visible.select do |value|
|
360
|
+
value.nil? or value
|
361
|
+
end
|
362
|
+
end
|
363
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
364
|
+
count visible
|
365
|
+
0 2 false
|
366
|
+
1 32 false
|
367
|
+
TABLE
|
368
|
+
end
|
369
|
+
|
370
|
+
test("reject") do
|
371
|
+
sliced_table = @table.slice do |slicer|
|
372
|
+
slicer.visible.reject do |value|
|
373
|
+
value.nil? or value
|
374
|
+
end
|
375
|
+
end
|
376
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
377
|
+
count visible
|
378
|
+
0 2 false
|
379
|
+
1 32 false
|
380
|
+
TABLE
|
381
|
+
end
|
382
|
+
|
383
|
+
test("!reject") do
|
384
|
+
sliced_table = @table.slice do |slicer|
|
385
|
+
!slicer.visible.reject do |value|
|
386
|
+
value.nil? or value
|
387
|
+
end
|
388
|
+
end
|
389
|
+
assert_equal(<<-TABLE, sliced_table.to_s)
|
390
|
+
count visible
|
391
|
+
0 0
|
392
|
+
1 1 true
|
393
|
+
2 4
|
394
|
+
3 8 true
|
395
|
+
4 16 true
|
396
|
+
5 64
|
397
|
+
6
|
398
|
+
7 256 true
|
399
|
+
TABLE
|
400
|
+
end
|
401
|
+
end
|