1 /* PSPP - computes sample statistics.
2 Copyright (C) 2007, 2008, 2009 Free Software Foundation, Inc.
4 This program is free software; you can redistribute it and/or
5 modify it under the terms of the GNU General Public License as
6 published by the Free Software Foundation; either version 2 of the
7 License, or (at your option) any later version.
9 This program is distributed in the hope that it will be useful, but
10 WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software
16 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
29 #include <libpspp/message.h>
30 #include <libpspp/version.h>
31 #include <gl/xalloc.h>
32 #include <data/dictionary.h>
33 #include <data/case.h>
34 #include <data/casereader.h>
35 #include <data/variable.h>
36 #include <data/attributes.h>
37 #include <data/file-handle-def.h>
38 #include <data/sys-file-writer.h>
39 #include <data/sys-file-reader.h>
40 #include <data/value.h>
41 #include <data/vardict.h>
42 #include <data/value-labels.h>
43 #include <data/format.h>
44 #include <data/data-in.h>
45 #include <data/data-out.h>
48 typedef struct fmt_spec input_format ;
49 typedef struct fmt_spec output_format ;
52 /* A thin wrapper around sfm_writer */
57 /* A pointer to the writer. The writer is owned by the struct */
58 struct casewriter *writer;
60 /* A pointer to the dictionary. Owned externally */
61 const struct dictionary *dict;
63 /* The scalar containing the dictionary */
68 /* A thin wrapper around sfm_reader */
71 struct sfm_read_info opts;
73 /* A pointer to the reader. The reader is owned by the struct */
74 struct casereader *reader;
76 /* A pointer to the dictionary. */
77 struct dictionary *dict;
82 /* A message handler which writes messages to PSPP::errstr */
84 message_handler (const struct msg *m)
86 SV *errstr = get_sv("PSPP::errstr", TRUE);
87 sv_setpv (errstr, m->text);
91 sysfile_close (struct sysfile_info *sfi)
97 retval = casewriter_destroy (sfi->writer);
105 scalar_to_value (union value *val, SV *scalar, const struct variable *var)
107 if ( var_is_numeric (var))
109 if ( SvNOK (scalar) || SvIOK (scalar) )
110 val->f = SvNV (scalar);
117 const char *p = SvPV (scalar, len);
118 int width = var_get_width (var);
119 value_set_missing (val, width);
120 memcpy (value_str_rw (val, width), p, len);
126 value_to_scalar (const union value *val, const struct variable *var)
128 if ( var_is_numeric (var))
130 if ( var_is_value_missing (var, val, MV_SYSTEM))
131 return newSVpvn ("", 0);
133 return newSVnv (val->f);
137 int width = var_get_width (var);
138 return newSVpvn (value_str (val, width), width);
144 var_set_input_format (struct variable *v, input_format ip_fmt)
146 struct fmt_spec *if_copy = malloc (sizeof (*if_copy));
147 memcpy (if_copy, &ip_fmt, sizeof (ip_fmt));
148 var_attach_aux (v, if_copy, var_dtor_free);
152 make_value_from_scalar (union value *uv, SV *val, const struct variable *var)
154 value_init (uv, var_get_width (var));
155 scalar_to_value (uv, val, var);
161 MODULE = PSPP PACKAGE = PSPP
167 assert (0 == strcmp (ver, bare_version));
169 msg_init (NULL, message_handler);
170 settings_init (0, 0);
174 format_value (val, var)
179 const struct fmt_spec *fmt = var_get_print_format (var);
180 const struct dictionary *dict = var_get_vardict (var)->dict;
183 make_value_from_scalar (&uv, val, var);
184 fprintf (stderr, "Encoding is %s\n", dict_get_encoding (dict));
185 s = data_out (&uv, dict_get_encoding (dict), fmt);
186 value_destroy (&uv, var_get_width (var));
187 ret = newSVpv (s, fmt->w);
195 value_is_missing (val, var)
201 make_value_from_scalar (&uv, val, var);
202 ret = var_is_value_missing (var, &uv, MV_ANY);
203 value_destroy (&uv, var_get_width (var));
210 MODULE = PSPP PACKAGE = PSPP::Dict
215 RETVAL = dict_create ();
222 struct dictionary *dict
229 struct dictionary *dict
231 RETVAL = dict_get_var_cnt (dict);
236 set_label (dict, label)
237 struct dictionary *dict
240 dict_set_label (dict, label);
243 set_documents (dict, docs)
244 struct dictionary *dict
247 dict_set_documents (dict, docs);
251 add_document (dict, doc)
252 struct dictionary *dict
255 dict_add_document_line (dict, doc);
259 clear_documents (dict)
260 struct dictionary *dict
262 dict_clear_documents (dict);
266 set_weight (dict, var)
267 struct dictionary *dict
270 dict_set_weight (dict, var);
274 pxs_get_variable (dict, idx)
275 struct dictionary *dict
278 SV *errstr = get_sv("PSPP::errstr", TRUE);
279 sv_setpv (errstr, "");
280 if ( SvIV (idx) >= dict_get_var_cnt (dict))
282 sv_setpv (errstr, "The dictionary doesn't have that many variables.");
286 RETVAL = dict_get_var (dict, SvIV (idx));
292 pxs_get_var_by_name (dict, name)
293 struct dictionary *dict
296 SV *errstr = get_sv("PSPP::errstr", TRUE);
297 sv_setpv (errstr, "");
299 struct variable *var = dict_lookup_var (dict, name);
301 sv_setpv (errstr, "No such variable.");
307 MODULE = PSPP PACKAGE = PSPP::Var
311 pxs_dict_create_var (dict, name, ip_fmt)
312 struct dictionary * dict
316 SV *errstr = get_sv("PSPP::errstr", TRUE);
317 sv_setpv (errstr, "");
318 if ( ! var_is_plausible_name (name, false))
320 sv_setpv (errstr, "The variable name is not valid.");
324 struct fmt_spec op_fmt;
327 op_fmt = fmt_for_output_from_input (&ip_fmt);
328 v = dict_create_var (dict, name,
329 fmt_is_string (op_fmt.type) ? op_fmt.w : 0);
332 sv_setpv (errstr, "The variable could not be created (probably already exists).");
335 var_set_both_formats (v, &op_fmt);
336 var_set_input_format (v, ip_fmt);
343 set_missing_values (var, v1, ...)
344 struct variable *var;
351 croak ("No more than 3 missing values are permitted");
353 for (i = 0; i < items - 1; ++i)
354 scalar_to_value (&val[i], ST(i+1), var);
356 struct missing_values mv;
357 mv_init (&mv, var_get_width (var));
358 for (i = 0 ; i < items - 1; ++i )
359 mv_add_value (&mv, &val[i]);
360 var_set_missing_values (var, &mv);
364 set_label (var, label)
365 struct variable *var;
368 var_set_label (var, label);
372 clear_value_labels (var)
373 struct variable *var;
375 var_clear_value_labels (var);
378 pxs_set_write_format (var, fmt)
382 var_set_write_format (var, &fmt);
386 pxs_set_print_format (var, fmt)
390 var_set_print_format (var, &fmt);
393 pxs_set_output_format (var, fmt)
397 var_set_both_formats (var, &fmt);
401 add_value_label (var, key, label)
406 SV *errstr = get_sv("PSPP::errstr", TRUE);
407 sv_setpv (errstr, "");
409 union value the_value;
410 int width = var_get_width (var);
413 value_init (&the_value, width);
414 if ( var_is_numeric (var))
416 if ( ! looks_like_number (key))
418 sv_setpv (errstr, "Cannot add label with string key to a numeric variable");
419 value_destroy (&the_value, width);
422 the_value.f = SvNV (key);
426 value_copy_str_rpad (&the_value, width, SvPV_nolen(key), ' ');
428 ok = var_add_value_label (var, &the_value, label);
429 value_destroy (&the_value, width);
432 sv_setpv (errstr, "Something went wrong");
442 HV *attrhash = (HV *) sv_2mortal ((SV *) newHV());
444 struct attrset *as = var_get_attributes (var);
448 struct attrset_iterator iter;
449 struct attribute *attr;
451 for (attr = attrset_first (as, &iter);
453 attr = attrset_next (as, &iter))
456 const char *name = attribute_get_name (attr);
458 AV *values = newAV ();
460 for (i = 0 ; i < attribute_get_n_values (attr); ++i )
462 const char *value = attribute_get_value (attr, i);
463 av_push (values, newSVpv (value, 0));
466 hv_store (attrhash, name, strlen (name),
467 newRV_noinc ((SV*) values), 0);
471 RETVAL = newRV ((SV *) attrhash);
478 struct variable * var
480 RETVAL = var_get_name (var);
487 struct variable * var
489 RETVAL = var_get_label (var);
495 get_value_labels (var)
498 HV *labelhash = (HV *) sv_2mortal ((SV *) newHV());
499 const struct val_lab *vl;
500 struct val_labs_iterator *viter = NULL;
501 const struct val_labs *labels = var_get_value_labels (var);
505 for (vl = val_labs_first (labels);
507 vl = val_labs_next (labels, vl))
509 SV *sv = value_to_scalar (&vl->value, var);
511 const char *s = SvPV (sv, len);
512 hv_store (labelhash, s, len, newSVpv (val_lab_get_label (vl), 0), 0);
516 RETVAL = newRV ((SV *) labelhash);
522 MODULE = PSPP PACKAGE = PSPP::Sysfile
525 struct sysfile_info *
526 pxs_create_sysfile (name, dict_ref, opts_hr)
531 SV *dict_sv = SvRV (dict_ref);
532 struct dictionary *dict = (void *) SvIV (dict_sv);
533 struct sfm_write_options opts;
534 if (!SvROK (opts_hr))
536 opts = sfm_writer_default_options ();
540 HV *opt_h = (HV *) SvRV (opts_hr);
541 SV** readonly = hv_fetch(opt_h, "readonly", 8, 0);
542 SV** compress = hv_fetch(opt_h, "compress", 8, 0);
543 SV** version = hv_fetch(opt_h, "version", 7, 0);
545 opts.create_writeable = readonly ? ! SvIV (*readonly) : true;
546 opts.compress = compress ? SvIV (*compress) : false;
547 opts.version = version ? SvIV (*version) : 3 ;
550 struct file_handle *fh =
551 fh_create_file (NULL, name, fh_default_properties () );
552 struct sysfile_info *sfi = xmalloc (sizeof (*sfi));
553 sfi->writer = sfm_open_writer (fh, dict, opts);
556 sfi->dict_sv = dict_sv;
557 SvREFCNT_inc (sfi->dict_sv);
565 struct sysfile_info *sfi
567 RETVAL = sysfile_close (sfi);
573 struct sysfile_info *sfi
576 SvREFCNT_dec (sfi->dict_sv);
580 append_case (sfi, ccase)
581 struct sysfile_info *sfi
584 SV *errstr = get_sv("PSPP::errstr", TRUE);
585 sv_setpv (errstr, "");
586 if ( (!SvROK(ccase)))
592 AV *av_case = (AV*) SvRV (ccase);
594 const struct variable **vv;
599 if ( av_len (av_case) >= dict_get_var_cnt (sfi->dict))
602 c = case_create (dict_get_proto (sfi->dict));
604 dict_get_vars (sfi->dict, &vv, &nv, 1u << DC_ORDINARY | 1u << DC_SYSTEM);
606 for (sv = av_shift (av_case); SvOK (sv); sv = av_shift (av_case))
608 const struct variable *v = vv[i++];
609 const struct fmt_spec *ifmt = var_get_aux (v);
611 /* If an input format has been set, then use it.
612 Otherwise just convert the raw value.
616 struct substring ss = ss_cstr (SvPV_nolen (sv));
617 if ( ! data_in (ss, LEGACY_NATIVE, ifmt->type, 0, 0, 0,
627 scalar_to_value (case_data_rw (c, v), sv, v);
631 /* The remaining variables must be sysmis or blank string */
632 while (i < dict_get_var_cnt (sfi->dict))
634 const struct variable *v = vv[i++];
635 union value *val = case_data_rw (c, v);
636 value_set_missing (val, var_get_width (v));
638 RETVAL = casewriter_write (sfi->writer, c);
647 MODULE = PSPP PACKAGE = PSPP::Reader
649 struct sysreader_info *
650 pxs_open_sysfile (name)
653 struct casereader *reader;
654 struct sysreader_info *sri = NULL;
655 struct file_handle *fh =
656 fh_create_file (NULL, name, fh_default_properties () );
658 sri = xmalloc (sizeof (*sri));
659 sri->reader = sfm_open_reader (fh, &sri->dict, &sri->opts);
661 if ( NULL == sri->reader)
673 pxs_get_dict (reader)
674 struct sysreader_info *reader;
676 RETVAL = reader->dict;
683 struct sysreader_info *sfr;
687 if (c = casereader_read (sfr->reader))
691 EXTEND (SP, dict_get_var_cnt (sfr->dict));
692 for (v = 0; v < dict_get_var_cnt (sfr->dict); ++v )
694 const struct variable *var = dict_get_var (sfr->dict, v);
695 const union value *val = case_data (c, var);
697 PUSHs (sv_2mortal (value_to_scalar (val, var)));