line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
#!/usr/bin/perl |
2
|
|
|
|
|
|
|
|
3
|
|
|
|
|
|
|
# |
4
|
|
|
|
|
|
|
# dbfilepivot.pm |
5
|
|
|
|
|
|
|
# Copyright (C) 2011-2016 by John Heidemann |
6
|
|
|
|
|
|
|
# |
7
|
|
|
|
|
|
|
# This program is distributed under terms of the GNU general |
8
|
|
|
|
|
|
|
# public license, version 2. See the file COPYING |
9
|
|
|
|
|
|
|
# in $dblibdir for details. |
10
|
|
|
|
|
|
|
# |
11
|
|
|
|
|
|
|
|
12
|
|
|
|
|
|
|
package Fsdb::Filter::dbfilepivot; |
13
|
|
|
|
|
|
|
|
14
|
|
|
|
|
|
|
=head1 NAME |
15
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
dbfilepivot - pivot a table, converting multiple rows into single wide row |
17
|
|
|
|
|
|
|
|
18
|
|
|
|
|
|
|
=head1 SYNOPSIS |
19
|
|
|
|
|
|
|
|
20
|
|
|
|
|
|
|
dbfilepivot [-e empty] -k KeyField -p PivotField [-v ValueField] |
21
|
|
|
|
|
|
|
|
22
|
|
|
|
|
|
|
=head1 DESCRIPTION |
23
|
|
|
|
|
|
|
|
24
|
|
|
|
|
|
|
Pivot a table, converting multiple rows corresponding to the |
25
|
|
|
|
|
|
|
same key into a single wide row. |
26
|
|
|
|
|
|
|
|
27
|
|
|
|
|
|
|
In a normalized database, one might have data with a schema like |
28
|
|
|
|
|
|
|
(id, attribute, value), |
29
|
|
|
|
|
|
|
but sometimes it's more convenient to see the data with a schema like |
30
|
|
|
|
|
|
|
(id, attribute1, attribute2). |
31
|
|
|
|
|
|
|
(For example, gnuplot's stacked histograms requires denormalized data.) |
32
|
|
|
|
|
|
|
Dbfilepivot converts the normalized format to the denormalized, |
33
|
|
|
|
|
|
|
but sometimes useful, format. |
34
|
|
|
|
|
|
|
Here the "id" is the key, the attribute is the "pivot", |
35
|
|
|
|
|
|
|
and the value is, well, the optional "value". |
36
|
|
|
|
|
|
|
|
37
|
|
|
|
|
|
|
An example is clearer. A gradebook usually looks like: |
38
|
|
|
|
|
|
|
|
39
|
|
|
|
|
|
|
#fsdb name hw_1 hw_2 hw_3 |
40
|
|
|
|
|
|
|
John 97 98 99 |
41
|
|
|
|
|
|
|
Paul - 80 82 |
42
|
|
|
|
|
|
|
|
43
|
|
|
|
|
|
|
but a properly normalized format would represent it as: |
44
|
|
|
|
|
|
|
|
45
|
|
|
|
|
|
|
#fsdb name hw score |
46
|
|
|
|
|
|
|
John 1 97 |
47
|
|
|
|
|
|
|
John 2 98 |
48
|
|
|
|
|
|
|
John 3 99 |
49
|
|
|
|
|
|
|
Paul 2 80 |
50
|
|
|
|
|
|
|
Paul 3 82 |
51
|
|
|
|
|
|
|
|
52
|
|
|
|
|
|
|
This tool converts the second form into the first, when used as |
53
|
|
|
|
|
|
|
|
54
|
|
|
|
|
|
|
dbfilepivot -k name -p hw -v score |
55
|
|
|
|
|
|
|
|
56
|
|
|
|
|
|
|
or |
57
|
|
|
|
|
|
|
|
58
|
|
|
|
|
|
|
dbfilepivot --possible-pivots='1 2 3' -k name -p hw -v score |
59
|
|
|
|
|
|
|
|
60
|
|
|
|
|
|
|
Here name is the I column that indicates which rows belong |
61
|
|
|
|
|
|
|
to the same entity, |
62
|
|
|
|
|
|
|
hw is the I column that will be indicate which column |
63
|
|
|
|
|
|
|
in the output is relevant, |
64
|
|
|
|
|
|
|
and score is the I that indicates what goes in the |
65
|
|
|
|
|
|
|
output. |
66
|
|
|
|
|
|
|
|
67
|
|
|
|
|
|
|
The pivot creates a new column C, C, etc. |
68
|
|
|
|
|
|
|
for each tag, the contents of the pivot field in the input. |
69
|
|
|
|
|
|
|
It then populates those new columns with the contents of the value field |
70
|
|
|
|
|
|
|
in the input. |
71
|
|
|
|
|
|
|
|
72
|
|
|
|
|
|
|
If no value column is specified, then values are either empty or 1. |
73
|
|
|
|
|
|
|
|
74
|
|
|
|
|
|
|
Dbfilepivot assumes all lines with the same key are adjacent |
75
|
|
|
|
|
|
|
in the input source, like L with the F<-S> option. |
76
|
|
|
|
|
|
|
To enforce this invariant, by default, it I input be sorted by key. |
77
|
|
|
|
|
|
|
|
78
|
|
|
|
|
|
|
There is no requirement that the pivot field be sorted (provided the key field is already sorted). |
79
|
|
|
|
|
|
|
|
80
|
|
|
|
|
|
|
By default, dbfilepivot makes two passes over its data |
81
|
|
|
|
|
|
|
and so requires temporary disk space equal to the input size. |
82
|
|
|
|
|
|
|
With the B<--possible-pivots> option, the user can specify pivots |
83
|
|
|
|
|
|
|
and skip the second pass and avoid temporary data storage. |
84
|
|
|
|
|
|
|
|
85
|
|
|
|
|
|
|
Memory usage is proportional to the number of unique pivot values. |
86
|
|
|
|
|
|
|
|
87
|
|
|
|
|
|
|
The inverse of this commend is L. |
88
|
|
|
|
|
|
|
|
89
|
|
|
|
|
|
|
|
90
|
|
|
|
|
|
|
=head1 OPTIONS |
91
|
|
|
|
|
|
|
|
92
|
|
|
|
|
|
|
=over 4 |
93
|
|
|
|
|
|
|
|
94
|
|
|
|
|
|
|
=item B<-k> or B<--key> KeyField |
95
|
|
|
|
|
|
|
|
96
|
|
|
|
|
|
|
specify which column is the key for grouping. |
97
|
|
|
|
|
|
|
Required (no default). |
98
|
|
|
|
|
|
|
|
99
|
|
|
|
|
|
|
=item B<-p> or B<--pivot> PivotField |
100
|
|
|
|
|
|
|
|
101
|
|
|
|
|
|
|
specify which column is the key to indicate which column in the output |
102
|
|
|
|
|
|
|
is relevant. |
103
|
|
|
|
|
|
|
Required (no default). |
104
|
|
|
|
|
|
|
|
105
|
|
|
|
|
|
|
=item B<-v> or B<--value> ValueField |
106
|
|
|
|
|
|
|
|
107
|
|
|
|
|
|
|
Specify which column is the value in the output. |
108
|
|
|
|
|
|
|
If none is given, 1 is used for the value. |
109
|
|
|
|
|
|
|
|
110
|
|
|
|
|
|
|
=item B<--possible-pivots PP> |
111
|
|
|
|
|
|
|
|
112
|
|
|
|
|
|
|
Specify all possible pivot values as PP, a whitespace-separated list. |
113
|
|
|
|
|
|
|
With this option, data is processed only once (not twice). |
114
|
|
|
|
|
|
|
|
115
|
|
|
|
|
|
|
=item B<-C S> or B<--element-separator S> |
116
|
|
|
|
|
|
|
|
117
|
|
|
|
|
|
|
Specify the separator I used to join the input's key column |
118
|
|
|
|
|
|
|
with its contents. |
119
|
|
|
|
|
|
|
(Defaults to a single underscore.) |
120
|
|
|
|
|
|
|
|
121
|
|
|
|
|
|
|
=item B<-e E> or B<--empty E> |
122
|
|
|
|
|
|
|
|
123
|
|
|
|
|
|
|
give value E as the value for empty (null) records |
124
|
|
|
|
|
|
|
|
125
|
|
|
|
|
|
|
=item B<-S> or B<--pre-sorted> |
126
|
|
|
|
|
|
|
|
127
|
|
|
|
|
|
|
Assume data is already grouped by key. |
128
|
|
|
|
|
|
|
Provided twice, it removes the validation of this assertion. |
129
|
|
|
|
|
|
|
By default, we sort by key. |
130
|
|
|
|
|
|
|
|
131
|
|
|
|
|
|
|
=item B<-T TmpDir> |
132
|
|
|
|
|
|
|
|
133
|
|
|
|
|
|
|
where to put tmp files. |
134
|
|
|
|
|
|
|
Also uses environment variable TMPDIR, if -T is |
135
|
|
|
|
|
|
|
not specified. |
136
|
|
|
|
|
|
|
Default is /tmp. |
137
|
|
|
|
|
|
|
|
138
|
|
|
|
|
|
|
=back |
139
|
|
|
|
|
|
|
|
140
|
|
|
|
|
|
|
=for comment |
141
|
|
|
|
|
|
|
begin_standard_fsdb_options |
142
|
|
|
|
|
|
|
|
143
|
|
|
|
|
|
|
This module also supports the standard fsdb options: |
144
|
|
|
|
|
|
|
|
145
|
|
|
|
|
|
|
=over 4 |
146
|
|
|
|
|
|
|
|
147
|
|
|
|
|
|
|
=item B<-d> |
148
|
|
|
|
|
|
|
|
149
|
|
|
|
|
|
|
Enable debugging output. |
150
|
|
|
|
|
|
|
|
151
|
|
|
|
|
|
|
=item B<-i> or B<--input> InputSource |
152
|
|
|
|
|
|
|
|
153
|
|
|
|
|
|
|
Read from InputSource, typically a file name, or C<-> for standard input, |
154
|
|
|
|
|
|
|
or (if in Perl) a IO::Handle, Fsdb::IO or Fsdb::BoundedQueue objects. |
155
|
|
|
|
|
|
|
|
156
|
|
|
|
|
|
|
=item B<-o> or B<--output> OutputDestination |
157
|
|
|
|
|
|
|
|
158
|
|
|
|
|
|
|
Write to OutputDestination, typically a file name, or C<-> for standard output, |
159
|
|
|
|
|
|
|
or (if in Perl) a IO::Handle, Fsdb::IO or Fsdb::BoundedQueue objects. |
160
|
|
|
|
|
|
|
|
161
|
|
|
|
|
|
|
=item B<--autorun> or B<--noautorun> |
162
|
|
|
|
|
|
|
|
163
|
|
|
|
|
|
|
By default, programs process automatically, |
164
|
|
|
|
|
|
|
but Fsdb::Filter objects in Perl do not run until you invoke |
165
|
|
|
|
|
|
|
the run() method. |
166
|
|
|
|
|
|
|
The C<--(no)autorun> option controls that behavior within Perl. |
167
|
|
|
|
|
|
|
|
168
|
|
|
|
|
|
|
=item B<--help> |
169
|
|
|
|
|
|
|
|
170
|
|
|
|
|
|
|
Show help. |
171
|
|
|
|
|
|
|
|
172
|
|
|
|
|
|
|
=item B<--man> |
173
|
|
|
|
|
|
|
|
174
|
|
|
|
|
|
|
Show full manual. |
175
|
|
|
|
|
|
|
|
176
|
|
|
|
|
|
|
=back |
177
|
|
|
|
|
|
|
|
178
|
|
|
|
|
|
|
=for comment |
179
|
|
|
|
|
|
|
end_standard_fsdb_options |
180
|
|
|
|
|
|
|
|
181
|
|
|
|
|
|
|
|
182
|
|
|
|
|
|
|
=head1 SAMPLE USAGE |
183
|
|
|
|
|
|
|
|
184
|
|
|
|
|
|
|
=head2 Input: |
185
|
|
|
|
|
|
|
|
186
|
|
|
|
|
|
|
#fsdb name hw score |
187
|
|
|
|
|
|
|
John 1 97 |
188
|
|
|
|
|
|
|
John 2 98 |
189
|
|
|
|
|
|
|
John 3 99 |
190
|
|
|
|
|
|
|
Paul 2 80 |
191
|
|
|
|
|
|
|
Paul 3 82 |
192
|
|
|
|
|
|
|
|
193
|
|
|
|
|
|
|
=head2 Command: |
194
|
|
|
|
|
|
|
|
195
|
|
|
|
|
|
|
cat data.fsdb | dbfilepivot -k name -p hw -v score |
196
|
|
|
|
|
|
|
|
197
|
|
|
|
|
|
|
=head2 Output: |
198
|
|
|
|
|
|
|
|
199
|
|
|
|
|
|
|
#fsdb name hw_1 hw_2 hw_3 |
200
|
|
|
|
|
|
|
John 97 98 99 |
201
|
|
|
|
|
|
|
Paul - 80 82 |
202
|
|
|
|
|
|
|
# | dbfilepivot -k name -p hw -v score |
203
|
|
|
|
|
|
|
|
204
|
|
|
|
|
|
|
=head1 SEE ALSO |
205
|
|
|
|
|
|
|
|
206
|
|
|
|
|
|
|
L. |
207
|
|
|
|
|
|
|
L. |
208
|
|
|
|
|
|
|
L. |
209
|
|
|
|
|
|
|
L. |
210
|
|
|
|
|
|
|
|
211
|
|
|
|
|
|
|
|
212
|
|
|
|
|
|
|
=head1 CLASS FUNCTIONS |
213
|
|
|
|
|
|
|
|
214
|
|
|
|
|
|
|
=cut |
215
|
|
|
|
|
|
|
|
216
|
|
|
|
|
|
|
@ISA = qw(Fsdb::Filter); |
217
|
|
|
|
|
|
|
$VERSION = 2.0; |
218
|
|
|
|
|
|
|
|
219
|
1
|
|
|
1
|
|
5667
|
use strict; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
23
|
|
220
|
1
|
|
|
1
|
|
4
|
use Pod::Usage; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
63
|
|
221
|
1
|
|
|
1
|
|
4
|
use Carp; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
37
|
|
222
|
|
|
|
|
|
|
|
223
|
1
|
|
|
1
|
|
4
|
use Fsdb::Filter; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
14
|
|
224
|
1
|
|
|
1
|
|
3
|
use Fsdb::IO::Reader; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
20
|
|
225
|
1
|
|
|
1
|
|
3
|
use Fsdb::IO::Writer; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
13
|
|
226
|
1
|
|
|
1
|
|
4
|
use Fsdb::IO::Replayable; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
22
|
|
227
|
1
|
|
|
1
|
|
4
|
use Fsdb::Filter::dbpipeline qw(dbpipeline_filter dbsort); |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
1267
|
|
228
|
|
|
|
|
|
|
|
229
|
|
|
|
|
|
|
|
230
|
|
|
|
|
|
|
=head2 new |
231
|
|
|
|
|
|
|
|
232
|
|
|
|
|
|
|
$filter = new Fsdb::Filter::dbfilepivot(@arguments); |
233
|
|
|
|
|
|
|
|
234
|
|
|
|
|
|
|
Create a new dbfilepivot object, taking command-line arguments. |
235
|
|
|
|
|
|
|
|
236
|
|
|
|
|
|
|
=cut |
237
|
|
|
|
|
|
|
|
238
|
|
|
|
|
|
|
sub new ($@) { |
239
|
0
|
|
|
0
|
1
|
|
my $class = shift @_; |
240
|
0
|
|
|
|
|
|
my $self = $class->SUPER::new(@_); |
241
|
0
|
|
|
|
|
|
bless $self, $class; |
242
|
0
|
|
|
|
|
|
$self->set_defaults; |
243
|
0
|
|
|
|
|
|
$self->parse_options(@_); |
244
|
0
|
|
|
|
|
|
$self->SUPER::post_new(); |
245
|
0
|
|
|
|
|
|
return $self; |
246
|
|
|
|
|
|
|
} |
247
|
|
|
|
|
|
|
|
248
|
|
|
|
|
|
|
|
249
|
|
|
|
|
|
|
=head2 set_defaults |
250
|
|
|
|
|
|
|
|
251
|
|
|
|
|
|
|
$filter->set_defaults(); |
252
|
|
|
|
|
|
|
|
253
|
|
|
|
|
|
|
Internal: set up defaults. |
254
|
|
|
|
|
|
|
|
255
|
|
|
|
|
|
|
=cut |
256
|
|
|
|
|
|
|
|
257
|
|
|
|
|
|
|
sub set_defaults ($) { |
258
|
0
|
|
|
0
|
1
|
|
my($self) = @_; |
259
|
0
|
|
|
|
|
|
$self->SUPER::set_defaults(); |
260
|
0
|
|
|
|
|
|
$self->{_elem_separator} = '_'; |
261
|
0
|
0
|
|
|
|
|
$self->{_tmpdir} = defined($ENV{'TMPDIR'}) ? $ENV{'TMPDIR'} : "/tmp"; |
262
|
0
|
|
|
|
|
|
$self->{_key_column} = undef; |
263
|
0
|
|
|
|
|
|
$self->{_pivot_column} = undef; |
264
|
0
|
|
|
|
|
|
$self->{_value_column} = undef; |
265
|
0
|
|
|
|
|
|
$self->{_pre_sorted} = 0; |
266
|
0
|
|
|
|
|
|
$self->{_sort_order} = undef; |
267
|
0
|
|
|
|
|
|
$self->{_sort_as_numeric} = undef; |
268
|
0
|
|
|
|
|
|
$self->{_possible_pivots} = undef; |
269
|
|
|
|
|
|
|
} |
270
|
|
|
|
|
|
|
|
271
|
|
|
|
|
|
|
=head2 parse_options |
272
|
|
|
|
|
|
|
|
273
|
|
|
|
|
|
|
$filter->parse_options(@ARGV); |
274
|
|
|
|
|
|
|
|
275
|
|
|
|
|
|
|
Internal: parse command-line arguments. |
276
|
|
|
|
|
|
|
|
277
|
|
|
|
|
|
|
=cut |
278
|
|
|
|
|
|
|
|
279
|
|
|
|
|
|
|
sub parse_options ($@) { |
280
|
0
|
|
|
0
|
1
|
|
my $self = shift @_; |
281
|
|
|
|
|
|
|
|
282
|
0
|
|
|
|
|
|
my(@argv) = @_; |
283
|
|
|
|
|
|
|
$self->get_options( |
284
|
|
|
|
|
|
|
\@argv, |
285
|
0
|
|
|
0
|
|
|
'help|?' => sub { pod2usage(1); }, |
286
|
0
|
|
|
0
|
|
|
'man' => sub { pod2usage(-verbose => 2); }, |
287
|
|
|
|
|
|
|
'autorun!' => \$self->{_autorun}, |
288
|
|
|
|
|
|
|
'close!' => \$self->{_close}, |
289
|
|
|
|
|
|
|
'C|element-separator=s' => \$self->{_elem_separator}, |
290
|
|
|
|
|
|
|
'd|debug+' => \$self->{_debug}, |
291
|
|
|
|
|
|
|
'e|empty=s' => \$self->{_empty}, |
292
|
0
|
|
|
0
|
|
|
'i|input=s' => sub { $self->parse_io_option('input', @_); }, |
293
|
|
|
|
|
|
|
'k|key=s' => \$self->{_key_column}, |
294
|
|
|
|
|
|
|
'log!' => \$self->{_logprog}, |
295
|
0
|
|
|
0
|
|
|
'o|output=s' => sub { $self->parse_io_option('output', @_); }, |
296
|
|
|
|
|
|
|
'p|pivot=s' => \$self->{_pivot_column}, |
297
|
|
|
|
|
|
|
'possible-pivots=s' => \$self->{_possible_pivots}, |
298
|
|
|
|
|
|
|
'S|pre-sorted+' => \$self->{_pre_sorted}, |
299
|
|
|
|
|
|
|
'T|tmpdir|tempdir=s' => \$self->{_tmpdir}, |
300
|
|
|
|
|
|
|
'v|value=s' => \$self->{_value_column}, |
301
|
|
|
|
|
|
|
# sort key options: |
302
|
0
|
|
|
0
|
|
|
'n|numeric' => sub { $self->{_sort_as_numeric} = 1; }, |
303
|
0
|
|
|
0
|
|
|
'N|lexical' => sub { $self->{_sort_as_numeric} = undef; }, |
304
|
0
|
|
|
0
|
|
|
'r|descending' => sub { $self->{_sort_order} = -1; }, |
305
|
0
|
|
|
0
|
|
|
'R|ascending' => sub { $self->{_sort_order} = 1; }, |
306
|
0
|
0
|
|
|
|
|
) or pod2usage(2); |
307
|
0
|
0
|
|
|
|
|
pod2usage(2) if ($#argv != -1); |
308
|
|
|
|
|
|
|
} |
309
|
|
|
|
|
|
|
|
310
|
|
|
|
|
|
|
=head2 _find_possible_pivots |
311
|
|
|
|
|
|
|
|
312
|
|
|
|
|
|
|
$filter->_find_possible_pivots(); |
313
|
|
|
|
|
|
|
|
314
|
|
|
|
|
|
|
Internal: scan input data to find all possible pivot values. |
315
|
|
|
|
|
|
|
|
316
|
|
|
|
|
|
|
Returns npivots, pivots_aref. |
317
|
|
|
|
|
|
|
|
318
|
|
|
|
|
|
|
=cut |
319
|
|
|
|
|
|
|
|
320
|
|
|
|
|
|
|
sub _find_possible_pivots($) { |
321
|
0
|
|
|
0
|
|
|
my($self) = @_; |
322
|
|
|
|
|
|
|
|
323
|
|
|
|
|
|
|
# |
324
|
|
|
|
|
|
|
# Read the data to find all possible pivots, |
325
|
|
|
|
|
|
|
# saving a copy as we go. |
326
|
|
|
|
|
|
|
# |
327
|
0
|
|
|
|
|
|
$self->{_replayable} = new Fsdb::IO::Replayable(-writer_args => [ -clone => $self->{_in} ], -reader_args => [ -comment_handler => $self->create_pass_comments_sub ]); |
328
|
0
|
|
|
|
|
|
my $save_out = $self->{_replayable_writer} = $self->{_replayable}->writer; |
329
|
0
|
|
|
|
|
|
my $read_fastpath_sub = $self->{_in}->fastpath_sub(); |
330
|
0
|
|
|
|
|
|
my $save_write_fastpath_sub = $save_out->fastpath_sub; |
331
|
0
|
|
|
|
|
|
my $fref; |
332
|
|
|
|
|
|
|
my %pivots; |
333
|
0
|
|
|
|
|
|
my $npivots = 0; |
334
|
|
|
|
|
|
|
my $loop = q( |
335
|
|
|
|
|
|
|
# first pass: reading data to find all possible pivots |
336
|
|
|
|
|
|
|
while ($fref = &$read_fastpath_sub) { |
337
|
|
|
|
|
|
|
my $value = $fref->[) . $self->{_pivot_coli} . q@]; |
338
|
0
|
|
|
|
|
|
if ($value ne '@ . $self->{_empty} . q@') { |
339
|
|
|
|
|
|
|
$npivots++ if (!defined($pivots{$value})); |
340
|
|
|
|
|
|
|
$pivots{$value} = 1; |
341
|
|
|
|
|
|
|
}; |
342
|
|
|
|
|
|
|
&$save_write_fastpath_sub($fref); |
343
|
|
|
|
|
|
|
}; |
344
|
|
|
|
|
|
|
@; |
345
|
0
|
0
|
|
|
|
|
print $loop if ($self->{_debug}); |
346
|
0
|
|
|
|
|
|
eval $loop; |
347
|
0
|
0
|
|
|
|
|
$@ && croak $self->{_prog} . ": internal eval error: $@.\n"; |
348
|
|
|
|
|
|
|
|
349
|
0
|
0
|
|
|
|
|
if (defined($self->{_sorter_fred})) { |
350
|
0
|
|
|
|
|
|
$self->{_sorter_fred}->join(); |
351
|
0
|
|
|
|
|
|
$self->{_sorter_fred} = undef; |
352
|
|
|
|
|
|
|
}; |
353
|
|
|
|
|
|
|
|
354
|
0
|
|
|
|
|
|
$self->{_replayable}->close; |
355
|
|
|
|
|
|
|
|
356
|
0
|
|
|
|
|
|
my(@pivots) = keys %pivots; |
357
|
|
|
|
|
|
|
|
358
|
0
|
|
|
|
|
|
return ($npivots, \@pivots); |
359
|
|
|
|
|
|
|
} |
360
|
|
|
|
|
|
|
|
361
|
|
|
|
|
|
|
|
362
|
|
|
|
|
|
|
=head2 _given_possible_pivots |
363
|
|
|
|
|
|
|
|
364
|
|
|
|
|
|
|
$filter->_given_possible_pivots(); |
365
|
|
|
|
|
|
|
|
366
|
|
|
|
|
|
|
Internal: parse option of possible pivots. |
367
|
|
|
|
|
|
|
|
368
|
|
|
|
|
|
|
Returns npivots, pivots_href. |
369
|
|
|
|
|
|
|
|
370
|
|
|
|
|
|
|
=cut |
371
|
|
|
|
|
|
|
|
372
|
|
|
|
|
|
|
sub _given_possible_pivots($) { |
373
|
0
|
|
|
0
|
|
|
my($self) = @_; |
374
|
|
|
|
|
|
|
|
375
|
|
|
|
|
|
|
# |
376
|
|
|
|
|
|
|
# All possible pivots are given by the user. |
377
|
|
|
|
|
|
|
# |
378
|
0
|
|
|
|
|
|
my @pivots = split(/\s+/, $self->{_possible_pivots}); |
379
|
0
|
|
|
|
|
|
return ($#pivots + 1, \@pivots); |
380
|
|
|
|
|
|
|
} |
381
|
|
|
|
|
|
|
|
382
|
|
|
|
|
|
|
|
383
|
|
|
|
|
|
|
=head2 setup |
384
|
|
|
|
|
|
|
|
385
|
|
|
|
|
|
|
$filter->setup(); |
386
|
|
|
|
|
|
|
|
387
|
|
|
|
|
|
|
Internal: setup, parse headers. |
388
|
|
|
|
|
|
|
|
389
|
|
|
|
|
|
|
=cut |
390
|
|
|
|
|
|
|
|
391
|
|
|
|
|
|
|
sub setup ($) { |
392
|
0
|
|
|
0
|
1
|
|
my($self) = @_; |
393
|
|
|
|
|
|
|
|
394
|
|
|
|
|
|
|
croak $self->{_prog} . ": invalid empty value (single quote).\n" |
395
|
0
|
0
|
|
|
|
|
if ($self->{_empty} eq "'"); |
396
|
|
|
|
|
|
|
|
397
|
|
|
|
|
|
|
# |
398
|
|
|
|
|
|
|
# guarantee data is sorted |
399
|
|
|
|
|
|
|
# (swap reader if necessary) |
400
|
0
|
0
|
|
|
|
|
if ($self->{_pre_sorted}) { |
401
|
|
|
|
|
|
|
# pre-sorted, so just read it |
402
|
0
|
|
|
|
|
|
$self->finish_io_option('input', -comment_handler => $self->create_delay_comments_sub); |
403
|
0
|
|
|
|
|
|
$self->{_sorter_fred} = undef; |
404
|
|
|
|
|
|
|
} else { |
405
|
|
|
|
|
|
|
# not sorted, so sort it and read that |
406
|
0
|
|
|
|
|
|
my @sort_args = ('--nolog', $self->{_key_column}); |
407
|
0
|
0
|
|
|
|
|
unshift(@sort_args, '--descending') if ($self->{_sort_order} == -1); |
408
|
0
|
0
|
|
|
|
|
unshift(@sort_args, ($self->{_sort_as_numeric} ? '--numeric' : '--lexical')); |
409
|
0
|
|
|
|
|
|
my($new_reader, $new_fred) = dbpipeline_filter($self->{_input}, [-comment_handler => $self->create_delay_comments_sub], dbsort(@sort_args)); |
410
|
0
|
|
|
|
|
|
$self->{_pre_sorted_input} = $self->{_input}; |
411
|
0
|
|
|
|
|
|
$self->{_in} = $new_reader; |
412
|
0
|
|
|
|
|
|
$self->{_sorter_fred} = $new_fred; |
413
|
|
|
|
|
|
|
}; |
414
|
|
|
|
|
|
|
|
415
|
0
|
0
|
|
|
|
|
pod2usage(2) if (!defined($self->{_key_column})); |
416
|
0
|
|
|
|
|
|
$self->{_key_coli} = $self->{_in}->col_to_i($self->{_key_column}); |
417
|
|
|
|
|
|
|
croak $self->{_prog} . ": key column " . $self->{_key_column} . " is not in input stream.\n" |
418
|
0
|
0
|
|
|
|
|
if (!defined($self->{_key_coli})); |
419
|
|
|
|
|
|
|
|
420
|
0
|
0
|
|
|
|
|
pod2usage(2) if (!defined($self->{_pivot_column})); |
421
|
0
|
|
|
|
|
|
$self->{_pivot_coli} = $self->{_in}->col_to_i($self->{_pivot_column}); |
422
|
|
|
|
|
|
|
croak $self->{_prog} . ": pivot column " . $self->{_pivot_column} . " is not in input stream.\n" |
423
|
0
|
0
|
|
|
|
|
if (!defined($self->{_pivot_coli})); |
424
|
|
|
|
|
|
|
|
425
|
0
|
0
|
|
|
|
|
if (defined($self->{_value_column})) { |
426
|
0
|
|
|
|
|
|
$self->{_value_coli} = $self->{_in}->col_to_i($self->{_value_column}); |
427
|
|
|
|
|
|
|
croak $self->{_prog} . ": value column " . $self->{_value_column} . " is not in input stream.\n" |
428
|
0
|
0
|
|
|
|
|
if (!defined($self->{_value_coli})); |
429
|
|
|
|
|
|
|
}; |
430
|
|
|
|
|
|
|
|
431
|
0
|
0
|
|
|
|
|
my($npivots, $pivots_aref) = (defined($self->{_possible_pivots}) ? $self->_given_possible_pivots() : $self->_find_possible_pivots()); |
432
|
0
|
0
|
|
|
|
|
croak $self->{_prog} . ": no input data or pivots\n" |
433
|
|
|
|
|
|
|
if ($npivots == 0); |
434
|
|
|
|
|
|
|
|
435
|
|
|
|
|
|
|
# |
436
|
|
|
|
|
|
|
# Now that we know the pivots, make the new columns. |
437
|
|
|
|
|
|
|
# |
438
|
|
|
|
|
|
|
# kill the old pivot column, and value if given. |
439
|
|
|
|
|
|
|
my @new_cols = grep(!($_ eq $self->{_pivot_column} || |
440
|
|
|
|
|
|
|
(defined($self->{_value_column}) && $_ eq $self->{_value_column})), |
441
|
0
|
|
0
|
|
|
|
@{$self->{_in}->cols}); |
|
0
|
|
|
|
|
|
|
442
|
0
|
|
|
|
|
|
$self->finish_io_option('output', -clone => $self->{_in}, -cols => \@new_cols, -outputheader => 'delay'); |
443
|
0
|
|
|
|
|
|
my %tag_colis; |
444
|
|
|
|
|
|
|
my %new_columns; |
445
|
0
|
|
|
|
|
|
foreach (sort @$pivots_aref) { |
446
|
|
|
|
|
|
|
# xxx: could try to sort numerically if all pivots are numbers |
447
|
0
|
|
|
|
|
|
my $new_column = $self->{_pivot_column} . $self->{_elem_separator} . $_; |
448
|
0
|
|
|
|
|
|
$new_columns{$new_column} = 1; |
449
|
|
|
|
|
|
|
$self->{_out}->col_create($new_column) |
450
|
0
|
0
|
|
|
|
|
or croak $self->{_prog} . ": cannot create column $new_column (maybe it already existed?)\n"; |
451
|
0
|
|
|
|
|
|
$tag_colis{$_} = $self->{_out}->col_to_i($new_column); |
452
|
|
|
|
|
|
|
}; |
453
|
0
|
|
|
|
|
|
$self->{_tag_colis_href} = \%tag_colis; |
454
|
|
|
|
|
|
|
# write the mapping code. |
455
|
0
|
|
|
|
|
|
my $old_mapping_code = ''; |
456
|
|
|
|
|
|
|
# first the old bits |
457
|
0
|
|
|
|
|
|
foreach (@{$self->{_in}->cols}) { |
|
0
|
|
|
|
|
|
|
458
|
|
|
|
|
|
|
next if ($_ eq $self->{_pivot_column} || |
459
|
0
|
0
|
0
|
|
|
|
(defined($self->{_value_column}) && $_ eq $self->{_value_column})); |
|
|
|
0
|
|
|
|
|
460
|
|
|
|
|
|
|
$old_mapping_code .= '$nf[' . $self->{_out}->col_to_i($_) . '] = ' . |
461
|
0
|
|
|
|
|
|
'$fref->[' . $self->{_in}->col_to_i($_) . '];' . "\n"; |
462
|
|
|
|
|
|
|
}; |
463
|
0
|
|
|
|
|
|
$self->{_old_mapping_code} = $old_mapping_code; |
464
|
|
|
|
|
|
|
# and initialize the new |
465
|
0
|
|
|
|
|
|
my $new_initialization_code = ''; |
466
|
0
|
|
|
|
|
|
foreach (sort keys %new_columns) { |
467
|
0
|
|
|
|
|
|
$new_initialization_code .= '$nf[' . $self->{_out}->col_to_i($_) . '] = ' . "\n"; |
468
|
|
|
|
|
|
|
}; |
469
|
0
|
|
|
|
|
|
$new_initialization_code .= "\t'" . $self->{_empty} . "';\n"; |
470
|
0
|
|
|
|
|
|
$self->{_new_initialization_code} = $new_initialization_code; |
471
|
|
|
|
|
|
|
} |
472
|
|
|
|
|
|
|
|
473
|
|
|
|
|
|
|
=head2 run |
474
|
|
|
|
|
|
|
|
475
|
|
|
|
|
|
|
$filter->run(); |
476
|
|
|
|
|
|
|
|
477
|
|
|
|
|
|
|
Internal: run over each rows. |
478
|
|
|
|
|
|
|
|
479
|
|
|
|
|
|
|
=cut |
480
|
|
|
|
|
|
|
sub run ($) { |
481
|
0
|
|
|
0
|
1
|
|
my($self) = @_; |
482
|
|
|
|
|
|
|
|
483
|
0
|
0
|
|
|
|
|
my $in_reader = (defined($self->{_replayable}) ? $self->{_replayable}->reader : $self->{_in}); |
484
|
0
|
|
|
|
|
|
my $read_fastpath_sub = $in_reader->fastpath_sub(); |
485
|
0
|
|
|
|
|
|
my $write_fastpath_sub = $self->{_out}->fastpath_sub(); |
486
|
|
|
|
|
|
|
|
487
|
|
|
|
|
|
|
# |
488
|
|
|
|
|
|
|
# Basic idea: mapreduce on the input |
489
|
|
|
|
|
|
|
# with a multikey aware reducer. |
490
|
|
|
|
|
|
|
# |
491
|
|
|
|
|
|
|
# We don't actually run mapreduce |
492
|
|
|
|
|
|
|
# because (sadly) it's easier to do it in-line |
493
|
|
|
|
|
|
|
# given we assume sorted input. |
494
|
|
|
|
|
|
|
# |
495
|
0
|
|
|
|
|
|
my $emit_nf_code = '&$write_fastpath_sub(\@nf);'; |
496
|
|
|
|
|
|
|
my $check_ordering_code = ' |
497
|
0
|
|
|
|
|
|
die "' . $self->{_prog} . q': keys $old_key and $new_key are out-of-order\n" if ($old_key gt $new_key); |
498
|
|
|
|
|
|
|
'; |
499
|
0
|
0
|
|
|
|
|
$check_ordering_code = '' if ($self->{_pre_sorted} > 1); |
500
|
0
|
0
|
|
|
|
|
my $value_value = (defined($self->{_value_column})) ? '$fref->[' . $self->{_value_coli} . ']' : '1'; |
501
|
0
|
|
|
|
|
|
my $tag_colis_href = $self->{_tag_colis_href}; |
502
|
|
|
|
|
|
|
my($loop) = q' |
503
|
|
|
|
|
|
|
{ |
504
|
|
|
|
|
|
|
my $old_key = undef; |
505
|
|
|
|
|
|
|
my $fref; |
506
|
|
|
|
|
|
|
my @nf; |
507
|
|
|
|
|
|
|
while ($fref = &$read_fastpath_sub()) { |
508
|
|
|
|
|
|
|
my $new_key = $fref->[' . $self->{_key_coli} . ']; |
509
|
|
|
|
|
|
|
if (!defined($old_key) || $old_key ne $new_key) { |
510
|
|
|
|
|
|
|
if (defined($old_key)) { |
511
|
|
|
|
|
|
|
' . $emit_nf_code . |
512
|
|
|
|
|
|
|
$check_ordering_code . ' |
513
|
|
|
|
|
|
|
}; |
514
|
|
|
|
|
|
|
' . $self->{_new_initialization_code} . ' |
515
|
|
|
|
|
|
|
' . $self->{_old_mapping_code} . ' |
516
|
|
|
|
|
|
|
$old_key = $new_key; |
517
|
|
|
|
|
|
|
}; |
518
|
0
|
|
|
|
|
|
my $pivot_value = $fref->[' . $self->{_pivot_coli} . ']; |
519
|
|
|
|
|
|
|
my $target_coli = $tag_colis_href->{$pivot_value}; |
520
|
|
|
|
|
|
|
die $self->{_prog} . ": unanticipated pivot value $pivot_value (forgot it in --possible-pivots?).\n" |
521
|
|
|
|
|
|
|
if (!defined($target_coli)); |
522
|
|
|
|
|
|
|
$nf[$target_coli] = ' . $value_value . '; |
523
|
|
|
|
|
|
|
}; |
524
|
|
|
|
|
|
|
if (defined($old_key)) { |
525
|
|
|
|
|
|
|
' . $emit_nf_code . " |
526
|
|
|
|
|
|
|
}; |
527
|
|
|
|
|
|
|
}\n"; |
528
|
0
|
0
|
|
|
|
|
print $loop if ($self->{_debug}); |
529
|
0
|
|
|
|
|
|
eval $loop; |
530
|
0
|
0
|
|
|
|
|
if ($@) { |
531
|
|
|
|
|
|
|
# propagate sort failure cleanly |
532
|
0
|
0
|
|
|
|
|
if ($@ =~ /^$self->{_prog}/) { |
533
|
0
|
|
|
|
|
|
croak "$@"; |
534
|
|
|
|
|
|
|
} else { |
535
|
0
|
|
|
|
|
|
croak $self->{_prog} . ": internal eval error: $@.\n"; |
536
|
|
|
|
|
|
|
}; |
537
|
|
|
|
|
|
|
}; |
538
|
|
|
|
|
|
|
|
539
|
|
|
|
|
|
|
# If single pass, we may need to collect this thread here. |
540
|
0
|
0
|
|
|
|
|
if (defined($self->{_sorter_fred})) { |
541
|
0
|
|
|
|
|
|
$self->{_sorter_fred}->join(); |
542
|
0
|
|
|
|
|
|
$self->{_sorter_fred} = undef; |
543
|
|
|
|
|
|
|
}; |
544
|
|
|
|
|
|
|
|
545
|
|
|
|
|
|
|
} |
546
|
|
|
|
|
|
|
|
547
|
|
|
|
|
|
|
|
548
|
|
|
|
|
|
|
=head1 AUTHOR and COPYRIGHT |
549
|
|
|
|
|
|
|
|
550
|
|
|
|
|
|
|
Copyright (C) 2011-2016 by John Heidemann |
551
|
|
|
|
|
|
|
|
552
|
|
|
|
|
|
|
This program is distributed under terms of the GNU general |
553
|
|
|
|
|
|
|
public license, version 2. See the file COPYING |
554
|
|
|
|
|
|
|
with the distribution for details. |
555
|
|
|
|
|
|
|
|
556
|
|
|
|
|
|
|
=cut |
557
|
|
|
|
|
|
|
|
558
|
|
|
|
|
|
|
1; |