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/lib/arrow/slicer.rb
ADDED
@@ -0,0 +1,391 @@
|
|
1
|
+
# Copyright 2017 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
|
+
module Arrow
|
16
|
+
class Slicer
|
17
|
+
def initialize(table)
|
18
|
+
@table = table
|
19
|
+
end
|
20
|
+
|
21
|
+
def [](column_name)
|
22
|
+
column = @table[column_name]
|
23
|
+
return nil if column.nil?
|
24
|
+
ColumnCondition.new(column)
|
25
|
+
end
|
26
|
+
|
27
|
+
def respond_to_missing?(name, include_private)
|
28
|
+
return true if self[name]
|
29
|
+
super
|
30
|
+
end
|
31
|
+
|
32
|
+
def method_missing(name, *args, &block)
|
33
|
+
if args.empty?
|
34
|
+
column_condition = self[name]
|
35
|
+
return column_condition if column_condition
|
36
|
+
end
|
37
|
+
super
|
38
|
+
end
|
39
|
+
|
40
|
+
class Condition
|
41
|
+
def evaluate
|
42
|
+
message = "Slicer::Condition must define \#evaluate: #{inspect}"
|
43
|
+
raise NotImplementedError.new(message)
|
44
|
+
end
|
45
|
+
|
46
|
+
def &(condition)
|
47
|
+
AndCondition.new(self, condition)
|
48
|
+
end
|
49
|
+
|
50
|
+
def |(condition)
|
51
|
+
OrCondition.new(self, condition)
|
52
|
+
end
|
53
|
+
|
54
|
+
def ^(condition)
|
55
|
+
XorCondition.new(self, condition)
|
56
|
+
end
|
57
|
+
end
|
58
|
+
|
59
|
+
class LogicalCondition < Condition
|
60
|
+
def initialize(condition1, condition2)
|
61
|
+
@condition1 = condition1
|
62
|
+
@condition2 = condition2
|
63
|
+
end
|
64
|
+
|
65
|
+
def evaluate
|
66
|
+
values1 = @condition1.evaluate.each
|
67
|
+
values2 = @condition2.evaluate.each
|
68
|
+
raw_array = []
|
69
|
+
begin
|
70
|
+
loop do
|
71
|
+
value1 = values1.next
|
72
|
+
value2 = values2.next
|
73
|
+
if value1.nil? or value2.nil?
|
74
|
+
raw_array << nil
|
75
|
+
else
|
76
|
+
raw_array << evaluate_value(value1, value2)
|
77
|
+
end
|
78
|
+
end
|
79
|
+
rescue StopIteration
|
80
|
+
end
|
81
|
+
BooleanArray.new(raw_array)
|
82
|
+
end
|
83
|
+
end
|
84
|
+
|
85
|
+
class AndCondition < LogicalCondition
|
86
|
+
private
|
87
|
+
def evaluate_value(value1, value2)
|
88
|
+
value1 and value2
|
89
|
+
end
|
90
|
+
end
|
91
|
+
|
92
|
+
class OrCondition < LogicalCondition
|
93
|
+
private
|
94
|
+
def evaluate_value(value1, value2)
|
95
|
+
value1 or value2
|
96
|
+
end
|
97
|
+
end
|
98
|
+
|
99
|
+
class XorCondition < LogicalCondition
|
100
|
+
private
|
101
|
+
def evaluate_value(value1, value2)
|
102
|
+
value1 ^ value2
|
103
|
+
end
|
104
|
+
end
|
105
|
+
|
106
|
+
class ColumnCondition < Condition
|
107
|
+
def initialize(column)
|
108
|
+
@column = column
|
109
|
+
end
|
110
|
+
|
111
|
+
def evaluate
|
112
|
+
data = @column.data
|
113
|
+
if data.n_chunks == 1
|
114
|
+
array = data.get_chunk(0)
|
115
|
+
if array.is_a?(BooleanArray)
|
116
|
+
array
|
117
|
+
else
|
118
|
+
array.cast(BooleanDataType.new)
|
119
|
+
end
|
120
|
+
else
|
121
|
+
raw_array = []
|
122
|
+
data.each_chunk do |chunk|
|
123
|
+
if chunk.is_a?(BooleanArray)
|
124
|
+
boolean_array = chunk
|
125
|
+
else
|
126
|
+
boolean_array = chunk.cast(BooleanDataType.new)
|
127
|
+
end
|
128
|
+
boolean_array.each do |value|
|
129
|
+
raw_array << value
|
130
|
+
end
|
131
|
+
end
|
132
|
+
BooleanArray.new(raw_array)
|
133
|
+
end
|
134
|
+
end
|
135
|
+
|
136
|
+
def !@
|
137
|
+
NotColumnCondition.new(@column)
|
138
|
+
end
|
139
|
+
|
140
|
+
def null?
|
141
|
+
self == nil
|
142
|
+
end
|
143
|
+
|
144
|
+
def ==(value)
|
145
|
+
EqualCondition.new(@column, value)
|
146
|
+
end
|
147
|
+
|
148
|
+
def !=(value)
|
149
|
+
NotEqualCondition.new(@column, value)
|
150
|
+
end
|
151
|
+
|
152
|
+
def <(value)
|
153
|
+
LessCondition.new(@column, value)
|
154
|
+
end
|
155
|
+
|
156
|
+
def <=(value)
|
157
|
+
LessEqualCondition.new(@column, value)
|
158
|
+
end
|
159
|
+
|
160
|
+
def >(value)
|
161
|
+
GreaterCondition.new(@column, value)
|
162
|
+
end
|
163
|
+
|
164
|
+
def >=(value)
|
165
|
+
GreaterEqualCondition.new(@column, value)
|
166
|
+
end
|
167
|
+
|
168
|
+
def select(&block)
|
169
|
+
SelectCondition.new(@column, block)
|
170
|
+
end
|
171
|
+
|
172
|
+
def reject(&block)
|
173
|
+
RejectCondition.new(@column, block)
|
174
|
+
end
|
175
|
+
end
|
176
|
+
|
177
|
+
class NotColumnCondition < Condition
|
178
|
+
def initialize(column)
|
179
|
+
@column = column
|
180
|
+
end
|
181
|
+
|
182
|
+
def evaluate
|
183
|
+
data = @column.data
|
184
|
+
raw_array = []
|
185
|
+
data.each_chunk do |chunk|
|
186
|
+
if chunk.is_a?(BooleanArray)
|
187
|
+
boolean_array = chunk
|
188
|
+
else
|
189
|
+
boolean_array = chunk.cast(BooleanDataType.new)
|
190
|
+
end
|
191
|
+
boolean_array.each do |value|
|
192
|
+
if value.nil?
|
193
|
+
raw_array << value
|
194
|
+
else
|
195
|
+
raw_array << !value
|
196
|
+
end
|
197
|
+
end
|
198
|
+
end
|
199
|
+
BooleanArray.new(raw_array)
|
200
|
+
end
|
201
|
+
|
202
|
+
def !@
|
203
|
+
ColumnCondition.new(@column)
|
204
|
+
end
|
205
|
+
end
|
206
|
+
|
207
|
+
class EqualCondition < Condition
|
208
|
+
def initialize(column, value)
|
209
|
+
@column = column
|
210
|
+
@value = value
|
211
|
+
end
|
212
|
+
|
213
|
+
def !@
|
214
|
+
NotEqualCondition.new(@column, @value)
|
215
|
+
end
|
216
|
+
|
217
|
+
def evaluate
|
218
|
+
case @value
|
219
|
+
when nil
|
220
|
+
raw_array = @column.collect(&:nil?)
|
221
|
+
BooleanArray.new(raw_array)
|
222
|
+
else
|
223
|
+
raw_array = @column.collect do |value|
|
224
|
+
if value.nil?
|
225
|
+
nil
|
226
|
+
else
|
227
|
+
@value == value
|
228
|
+
end
|
229
|
+
end
|
230
|
+
BooleanArray.new(raw_array)
|
231
|
+
end
|
232
|
+
end
|
233
|
+
end
|
234
|
+
|
235
|
+
class NotEqualCondition < Condition
|
236
|
+
def initialize(column, value)
|
237
|
+
@column = column
|
238
|
+
@value = value
|
239
|
+
end
|
240
|
+
|
241
|
+
def !@
|
242
|
+
EqualCondition.new(@column, @value)
|
243
|
+
end
|
244
|
+
|
245
|
+
def evaluate
|
246
|
+
case @value
|
247
|
+
when nil
|
248
|
+
raw_array = @column.collect do |value|
|
249
|
+
not value.nil?
|
250
|
+
end
|
251
|
+
BooleanArray.new(raw_array)
|
252
|
+
else
|
253
|
+
raw_array = @column.collect do |value|
|
254
|
+
if value.nil?
|
255
|
+
nil
|
256
|
+
else
|
257
|
+
@value != value
|
258
|
+
end
|
259
|
+
end
|
260
|
+
BooleanArray.new(raw_array)
|
261
|
+
end
|
262
|
+
end
|
263
|
+
end
|
264
|
+
|
265
|
+
class LessCondition < Condition
|
266
|
+
def initialize(column, value)
|
267
|
+
@column = column
|
268
|
+
@value = value
|
269
|
+
end
|
270
|
+
|
271
|
+
def !@
|
272
|
+
GreaterEqualCondition.new(@column, @value)
|
273
|
+
end
|
274
|
+
|
275
|
+
def evaluate
|
276
|
+
raw_array = @column.collect do |value|
|
277
|
+
if value.nil?
|
278
|
+
nil
|
279
|
+
else
|
280
|
+
@value > value
|
281
|
+
end
|
282
|
+
end
|
283
|
+
BooleanArray.new(raw_array)
|
284
|
+
end
|
285
|
+
end
|
286
|
+
|
287
|
+
class LessEqualCondition < Condition
|
288
|
+
def initialize(column, value)
|
289
|
+
@column = column
|
290
|
+
@value = value
|
291
|
+
end
|
292
|
+
|
293
|
+
def !@
|
294
|
+
GreaterCondition.new(@column, @value)
|
295
|
+
end
|
296
|
+
|
297
|
+
def evaluate
|
298
|
+
raw_array = @column.collect do |value|
|
299
|
+
if value.nil?
|
300
|
+
nil
|
301
|
+
else
|
302
|
+
@value >= value
|
303
|
+
end
|
304
|
+
end
|
305
|
+
BooleanArray.new(raw_array)
|
306
|
+
end
|
307
|
+
end
|
308
|
+
|
309
|
+
class GreaterCondition < Condition
|
310
|
+
def initialize(column, value)
|
311
|
+
@column = column
|
312
|
+
@value = value
|
313
|
+
end
|
314
|
+
|
315
|
+
def !@
|
316
|
+
LessEqualCondition.new(@column, @value)
|
317
|
+
end
|
318
|
+
|
319
|
+
def evaluate
|
320
|
+
raw_array = @column.collect do |value|
|
321
|
+
if value.nil?
|
322
|
+
nil
|
323
|
+
else
|
324
|
+
@value < value
|
325
|
+
end
|
326
|
+
end
|
327
|
+
BooleanArray.new(raw_array)
|
328
|
+
end
|
329
|
+
end
|
330
|
+
|
331
|
+
class GreaterEqualCondition < Condition
|
332
|
+
def initialize(column, value)
|
333
|
+
@column = column
|
334
|
+
@value = value
|
335
|
+
end
|
336
|
+
|
337
|
+
def !@
|
338
|
+
LessCondition.new(@column, @value)
|
339
|
+
end
|
340
|
+
|
341
|
+
def evaluate
|
342
|
+
raw_array = @column.collect do |value|
|
343
|
+
if value.nil?
|
344
|
+
nil
|
345
|
+
else
|
346
|
+
@value <= value
|
347
|
+
end
|
348
|
+
end
|
349
|
+
BooleanArray.new(raw_array)
|
350
|
+
end
|
351
|
+
end
|
352
|
+
|
353
|
+
class SelectCondition < Condition
|
354
|
+
def initialize(column, block)
|
355
|
+
@column = column
|
356
|
+
@block = block
|
357
|
+
end
|
358
|
+
|
359
|
+
def !@
|
360
|
+
RejectCondition.new(@column, @block)
|
361
|
+
end
|
362
|
+
|
363
|
+
def evaluate
|
364
|
+
BooleanArray.new(@column.collect(&@block))
|
365
|
+
end
|
366
|
+
end
|
367
|
+
|
368
|
+
class RejectCondition < Condition
|
369
|
+
def initialize(column, block)
|
370
|
+
@column = column
|
371
|
+
@block = block
|
372
|
+
end
|
373
|
+
|
374
|
+
def !@
|
375
|
+
SelectCondition.new(@column, @block)
|
376
|
+
end
|
377
|
+
|
378
|
+
def evaluate
|
379
|
+
raw_array = @column.collect do |value|
|
380
|
+
evaluated_value = @block.call(value)
|
381
|
+
if evaluated_value.nil?
|
382
|
+
nil
|
383
|
+
else
|
384
|
+
not evaluated_value
|
385
|
+
end
|
386
|
+
end
|
387
|
+
BooleanArray.new(raw_array)
|
388
|
+
end
|
389
|
+
end
|
390
|
+
end
|
391
|
+
end
|
@@ -0,0 +1,88 @@
|
|
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
|
+
require "time"
|
16
|
+
|
17
|
+
module Arrow
|
18
|
+
class TableFormatter
|
19
|
+
def initialize(table, options={})
|
20
|
+
@table = table
|
21
|
+
@options = options
|
22
|
+
end
|
23
|
+
|
24
|
+
def format
|
25
|
+
text = ""
|
26
|
+
columns = @table.columns
|
27
|
+
columns.each do |column|
|
28
|
+
text << "\t"
|
29
|
+
text << format_column_name(column)
|
30
|
+
end
|
31
|
+
text << "\n"
|
32
|
+
|
33
|
+
n_rows = @table.n_rows
|
34
|
+
return text if n_rows.zero?
|
35
|
+
|
36
|
+
border = @options[:border] || 10
|
37
|
+
n_digits = (Math.log10(n_rows) + 1).truncate
|
38
|
+
[border, n_rows].min.times do |i|
|
39
|
+
format_row(text, columns, i, n_digits)
|
40
|
+
end
|
41
|
+
return text if n_rows <= border
|
42
|
+
|
43
|
+
text << "...\n"
|
44
|
+
[border, n_rows - border].max.upto(n_rows - 1) do |i|
|
45
|
+
format_row(text, columns, i, n_digits)
|
46
|
+
end
|
47
|
+
|
48
|
+
text
|
49
|
+
end
|
50
|
+
|
51
|
+
private
|
52
|
+
FLOAT_N_DIGITS = 10
|
53
|
+
def format_column_name(column)
|
54
|
+
case column.data_type
|
55
|
+
when TimestampDataType
|
56
|
+
"%*s" % [Time.now.iso8601.size, column.name]
|
57
|
+
when FloatDataType, DoubleDataType
|
58
|
+
"%*s" % [FLOAT_N_DIGITS, column.name]
|
59
|
+
else
|
60
|
+
column.name
|
61
|
+
end
|
62
|
+
end
|
63
|
+
|
64
|
+
def format_row(text, columns, i, n_digits)
|
65
|
+
text << ("%*d" % [n_digits, i])
|
66
|
+
columns.each do |column|
|
67
|
+
text << "\t"
|
68
|
+
text << format_column_value(column, i)
|
69
|
+
end
|
70
|
+
text << "\n"
|
71
|
+
text
|
72
|
+
end
|
73
|
+
|
74
|
+
def format_column_value(column, i)
|
75
|
+
value = column[i]
|
76
|
+
case value
|
77
|
+
when Time
|
78
|
+
value.iso8601
|
79
|
+
when Float
|
80
|
+
"%*f" % [[column.name.size, FLOAT_N_DIGITS].max, value]
|
81
|
+
when Integer
|
82
|
+
"%*d" % [column.name.size, value]
|
83
|
+
else
|
84
|
+
"%-*s" % [column.name.size, value.to_s]
|
85
|
+
end
|
86
|
+
end
|
87
|
+
end
|
88
|
+
end
|