[37798a]: inst / @dataframe / subsref.m  Maximize  Restore  History

Download this file

454 lines (444 with data), 14.4 kB

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
function resu = subsref(df, S)
%# function resu = subsref(df, S)
%# This function returns a subpart of a dataframe. It is invoked when
%# calling df.field, df(value), or df{value}. In case of fields,
%# returns either the content of the container with the same name,
%# either the column with the same name, priority being given to the
%# container. In case of range, selection may occur on name or order
%# (not rowidx for rows). If the result is homogenous, it is
%# downclassed. In case an extra field is given, is it used to
%# determine the class of the return value. F.i.,
%# df(1, 2, 'dataframe')
%# does not return a scalar but a dataframe, keeping all the meta-information
%% Copyright (C) 2009-2010 Pascal Dupuis <Pascal.Dupuis@uclouvain.be>
%%
%% This file is part of Octave.
%%
%% Octave is free software; you can redistribute it and/or
%% modify it under the terms of the GNU General Public
%% License as published by the Free Software Foundation;
%% either version 2, or (at your option) any later version.
%%
%% Octave is distributed in the hope that it will be useful,
%% but WITHOUT ANY WARRANTY; without even the implied
%% warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
%% PURPOSE. See the GNU General Public License for more
%% details.
%%
%% You should have received a copy of the GNU General Public
%% License along with Octave; see the file COPYING. If not,
%% write to the Free Software Foundation, 59 Temple Place -
%% Suite 330, Boston, MA 02111-1307, USA.
%#
%# $Id: subsref.m 1034 2010-08-03 16:22:09Z dupuis $
%#
if strcmp(S(1).type, '.'), %# struct access
resu = struct(df); %# avoid recursive calls
if strfind(S(1).subs, '_'), %# its an internal field name
%# FIXME: this should only be called from class members and friends
resu = builtin('subsref', resu, S);
else
%# direct access through the column name
indi = strmatch(S(1).subs, resu._name{2});
if ~isempty(indi),
resu = df._data{indi}; %# extract colum;
if strcmp(df._type{indi}, 'char') ...
&& 1 == size(df._data{indi}, 2),
resu = char(resu);
endif
if length(S) > 1,
dummy = S(2:end); S = S(1);
switch dummy(1).type
case '()'
if isa(dummy(1).subs{1}, "char"),
[indr, nrow, dummy(1).subs{1}] = ...
df_name2idx(df._name{1}, dummy(1).subs{1}, df._cnt(1), 'row');
endif
resu = builtin('subsref', resu, dummy);
otherwise
error("Invalid column access");
endswitch
endif
else %# access of one of the internal fieldnames
dummy = S(2:end); S = S(1);
postop = ''; preop = '()';
%# translate the external to internal name
switch S(1).subs
case "rownames"
S(1).subs = "_name";
S(2).type = "{}"; S(2).subs{1}= 1;
postop = @(x) char(x);
case "colnames"
S(1).subs = "_name";
S(2).type = "{}"; S(2).subs{1}= 2;
postop = @(x) char(x);
case "rowcnt"
S(1).subs = "_cnt";
S(2).type = "()"; S(2).subs{1}= 1;
case "colcnt"
S(1).subs = "_cnt";
S(2).type = "()"; S(2).subs{1}= 2;
case "rowidx"
S(1).subs = "_ridx";
case "types" %# this one should be accessed as a matrix
S(1).subs = "_type"; preop = '{}';
endswitch
if !isempty(dummy),
if isa(dummy(1).subs{1}, "char"),
[indc, ncol, dummy(1).subs{1}] = ...
df_name2idx(df._name{2}, dummy(1).subs{1}, ...
df._cnt(2), 'column');
if isempty(indc),
endif
endif
if !strcmp(dummy(1).type, preop),
error("Invalid internal field name access");
endif
endif
S = [S dummy];
resu = builtin('subsref', resu, S);
if !isempty(postop),
resu = postop(resu);
endif
endif
endif
return
endif
%# disp('line 103 '); keyboard
while 1, %#global loop, to avoid recursive accesses
asked_output_type = '';
%# iterate over S, avoid strange construct as x()()(1:10, 1:4)
while 1,
if !strcmp(S(1).type, '()'),
disp(S); keyboard
error("Illegal dataframe dereferencing");
endif
if isempty(S(1).subs),
if isempty(asked_output_type),
asked_output_type = class(df);
endif
if length(S) <= 1,
if strcmp(asked_output_type, class(df))
resu = df; %# whole access without conversion
return
endif
nrow = df._cnt(1); indr = 1:nrow;
ncol = df._cnt(2); indc = 1:ncol;
break; %# no dimension specified -- select all, the
%# asked_output_type was set in a previous iteration
else
%# avoid recursive calls
S = S(2:end);
continue;
endif
elseif isa(S(1).subs{end}, 'char'),
switch(S(1).subs{end}) %# remove last argument of 'dataframe', 'cell'
case {class(df) 'cell'}
if !isempty(asked_output_type) ...
&& !strcmp(asked_output_type, class(df))
error("Required output type set more than once")
endif
asked_output_type = S(1).subs{end};
S(1).subs = S(1).subs(1:end-1);
if isempty(S(1).subs) && length(S) > 1,
S = S(2:end);
endif
continue; %# restart the parsing
case { 'df' }
if !isempty(asked_output_type) ...
&& !strcmp(asked_output_type, class(df))
error("Required output type set more than once")
endif
asked_output_type = 'dataframe';
S(1).subs = S(1).subs(1:end-1);
if isempty(S(1).subs) && length(S) > 1,
S = S(2:end);
endif
continue; %# restart the parsing
otherwise
endswitch
endif
%# generic access
if isempty(S(1).subs{1}),
error('subsref: first dimension empty ???');
endif
if length(S(1).subs) > 1,
if isempty(S(1).subs{2}),
error('subsref: second dimension empty ???');
endif
[indr, nrow, S(1).subs{1}] = ...
df_name2idx(df._name{1}, S(1).subs{1}, df._cnt(1), 'row');
if !isa(indr, 'char') && max(indr) > df._cnt(1),
error("Accessing dataframe past end of lines");
endif
[indc, ncol, S(1).subs{2}] = ...
df_name2idx(df._name{2}, S(1).subs{2}, df._cnt(2), 'column');
if max(indc) > df._cnt(2),
error("Accessing dataframe past end of columns");
endif
else
%# one single dim -- probably something like df(:)
switch class(S(1).subs{1})
case {'char'} %# one dimensional access, disallow it if not ':'
if strcmp(S(1).subs{1}, ':'),
nrow = df._cnt(1); indr = 1:nrow;
ncol = df._cnt(2); indc = 1:ncol;
fullindr = []; fullindc = [];
else
error(["Accessing through single dimension and name " ...
S(1).subs{1} " not allowed\n-- use variable(:, 'name') instead"]);
endif
otherwise
[fullindr, fullindc] = ind2sub(df._cnt, S(1).subs{1});
indr = unique(fullindr); indc = unique(fullindc);
nrow = length(indr); ncol = length(indc);
if !isempty(asked_output_type) && ncol > 1,
%# verify that the extracted values form a square matrix
dummy = zeros(indr(end), indc(end));
for indi = 1:ncol,
indj = find(fullindc == indc(indi));
dummy(fullindr(indj), indc(indi)) = 1;
endfor
dummy = dummy(indr(1):indr(end), indc(1):indc(end));
if any(any(dummy!= 1)),
error("Vector-like selection is not rectangular for the asked output type");
else
fullindr = []; fullindc = [];
endif
endif
endswitch
endif
break;
endwhile
if ~isempty(indr) && ~isempty(indc),
%# can the data be merged ?
output_type = df._type{indc(1)};
for indi = 2:ncol,
if ~strcmp(output_type, df._type{indc(indi)}),
%# mixed args -- falls back to type of parent container
output_type = class(df);
break;
endif
endfor
if any(strcmp({output_type, asked_output_type}, class(df))),
if 1 == length(S(1).subs),
if ncol > 1
if isempty(asked_output_type),
%# sanity check
error("Selected columns not compatible with cat() -- use 'cell' as output format");
endif
if strcmp(asked_output_type, class(df))
error("Vector-like access not implemented for 'dataframe' output format");
endif
endif
endif
endif
if !isempty(asked_output_type),
%# override the class of the return value
output_type = asked_output_type;
endif
indt = {};
if length(S(1).subs) > 1, %# access-as-matrix
if length(S(1).subs) > 2,
inds = S(1).subs{3};
if isa(inds, 'char'),
nseq = max(cellfun('size', df._data(indc), 2));
indt(1, 1:df._cnt(2)) = inds;
else
%# generate a specific index for each column
nseq = length(inds);
dummy = cellfun('size', df._data(indc), 2);
indt(1, 1:df._cnt(2)) = inds;
indt(1==dummy) = 1;
endif
elseif all(cellfun('isclass', S(1).subs, 'char')),
inds = ':'; indt(1, 1:df._cnt(2)) = inds;
nseq = max(cellfun('size', df._data(indc), 2));
else
inds = 1; indt(1, 1:df._cnt(2)) = inds; nseq = 1;
endif
endif
if strcmp(output_type, class(df)),
%# export the result as a dataframe
resu = dataframe([]);
resu._cnt(1) = nrow; resu._cnt(2) = ncol;
for indi = 1:ncol,
resu._data{indi} = df._data{indc(indi)}...
(indr, indt{indc(indi)});
resu._name{2}(indi, 1) = df._name{2}(indc(indi));
resu._over{2}(1, indi) = df._over{2}(indc(indi));
resu._type{indi} = df._type{indc(indi)};
endfor
resu._ridx = df._ridx(indr, inds);
if length(df._name{1}) >= max(indr),
resu._name{1}(1:nrow, 1) = df._name{1}(indr);
resu._over{1}(1, 1:nrow) = df._over{1}(indr);
endif
dummy = sum(cellfun('size', resu._data, 2));
if (dummy > resu._cnt(2)),
resu._cnt(3) = dummy;
else
resu._cnt = resu._cnt(1:2);
endif
if length(S) > 1, %# perform further access, if required
df = resu;
S = S(2:end); %# avoid recursive calls
continue;
endif
return;
elseif strcmp(output_type, 'cell'),
%# export the result as a cell array
resu = cell(2+nrow, 2+ncol); resu(1:end, 1:2) = {''};
resu(2, 3:end) = df._type(indc); %column type
for indi = 1:ncol,
resu{1, 2+indi} = df._name{2}{indc(indi)}; % column name
switch df._type{indc(indi)} % cell content
case {'char' }
dummy = cellstr(df._data{indc(indi)}(indr, :));
resu(3:end, 2+indi) = dummy;
otherwise
dummy = df._data{indc(indi)}(indr, :);
resu(3:end, 2+indi) = mat2cell(dummy, ones(nrow, 1), 1);
endswitch
endfor
resu(3:end, 1) = mat2cell(df._ridx(indr), ones(nrow, 1), 1);
if length(df._name{1}) >= max(indr),
resu(3:end, 2) = df._name{1}{indr};
endif
if length(S) > 1, %# perform further access, if required
if length(S(2).subs) != 1,
%# normal, two-dimensionnal access
%# apply the selection on the zone containing the data
dummy = S;
if !isempty(dummy(2).subs),
dummy(2).subs{2} = ':';
endif
resuf = cat(2, ...
%# reselect indexes
builtin('subsref', resu(3:end, 1), dummy(2:end)), ...
%# reselect rownames
builtin('subsref', resu(3:end, 2), dummy(2:end)), ...
%# extract - reorder - whatever
builtin('subsref', resu(3:end, 3:end), S(2:end)) ...
);
dummy = S;
if !isempty(dummy(2).subs),
dummy(2).subs{1} = [1 2];
endif
resuf = cat(1, ...
%# reselect column names and types
[cell(2, 2) builtin('subsref', resu(1:2, 3:end), ...
dummy(2:end))], ...
resuf ...
);
resuf(1:2, 1:2) = {''}; resu = resuf;
else
%# do not expect this to produce meaningfull results, but you
%# asked it -- you got it
resu = builtin('subsref', resu(:), S(2:end));
if !isa(S(2).subs{1}, 'char') ...
&& size(S(2).subs{1}, 2) > 1,
resu = resu.';
endif
endif
elseif 1 == length(S(1).subs),
resu = resu(:);
if !isa(S(1).subs{1}, 'char') ...
&& size(S(1).subs{1}, 2) > 1,
resu = resu.';
endif
endif
return; %# no more iteration required
else
%# export the result as a vector/matrix
%# rules
%# x(:, :, :) returns a 3D matrix
%# x(:, n:m, :) returns a 3D matrix
%# x(:, :) returns a horzcat of the third dimension
%# x(:, n:m) select only the first sequence
%# x(:) returns a vertcat of the columns of x(:, :)
%# disp('line 329 '); keyboard
if length(S(1).subs) > 1, %# access-as-matrix
dummy = df._data{indc(1)};
%# disp('line 343'); keyboard
if ncol > 1,
%# dynamic allocation with the final type
if size(dummy, 2)> 1,
%# disp('line 347: FIXME '); keyboard
resu = repmat(dummy(indr, inds(1)), [1 ncol nseq]);
%# "turn" the extracted matrix
resu(:, 1, :) = dummy(indr, inds);
else
resu = repmat(dummy(indr), [1 ncol nseq]);
endif
for indi = 2:ncol,
dummy = df._data{indc(indi)};
if 1 == size(dummy, 2),
if 1 == nseq,
resu(:, indi) = dummy(indr);
else
resu(:, indi, :) = repmat(dummy(indr), [1 1 nseq]);
endif
else
%# "turn" the extracted matrix
resu(:, indi, :) = dummy(indr, inds);
endif
endfor
else
if strcmp(df._type{indc(1)}, 'char'),
resu = char(dummy(indr, inds));
else
resu = dummy(indr, inds);
endif
endif
if 2 == length(S(1).subs) ...
&& all(cellfun('isclass', S(1).subs, 'char')),
resu = reshape(resu, nrow, ncol*nseq);
endif
else
if !isempty(fullindr),
indj = find(fullindc == indc(1));
resu = df._data{indc(1)}(fullindr(indj));
switch df._type{indc(1)}
case {'char'}
for indi = 2:ncol,
indj = find(fullindc == indc(indi));
resu = char(resu, df._data{indc(indi)}(fullindr(indj)));
endfor
otherwise
for indi = 2:ncol,
indj = find(fullindc == indc(indi));
resu = cat(1, resu, df._data{indc(indi)}(fullindr(indj)));
endfor
endswitch
else
disp('line 389')
resu = df._data{indc(1)}(indr);
switch df._type{indc(1)}
case {'char'}
for indi = 2:ncol,
resu = char(resu, df._data{indc(indi)}(indr));
endfor
otherwise
for indi = 2:ncol,
resu = cat(1, resu, df._data{indc(indi)}(indr));
endfor
endswitch
endif
if !isa(S(1).subs{1}, 'char') ...
&& size(S(1).subs{1}, 2) > 1,
resu = resu.';
endif
endif
if length(S) > 1, %# perform further access, if required
resu = builtin('subsref', resu, S(2:end));
endif
endif
return; %# no more iteration required
else
resu = dataframe([]); %# returns empty object
endif
disp("line 343 !?!"); keyboard
return
endwhile
endfunction

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

JavaScript is required for this form.





No, thanks