1 /* PSPP - computes sample statistics.
2 Copyright (C) 1997-9, 2000 Free Software Foundation, Inc.
3 Written by Ben Pfaff <blp@gnu.org>.
5 This program is free software; you can redistribute it and/or
6 modify it under the terms of the GNU General Public License as
7 published by the Free Software Foundation; either version 2 of the
8 License, or (at your option) any later version.
10 This program is distributed in the hope that it will be useful, but
11 WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with this program; if not, write to the Free Software
17 Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
20 /* AIX requires this to be the first thing in the file. */
23 #define alloca __builtin_alloca
31 #ifndef alloca /* predefined by HP cc +Olibcalls */
44 #include <unistd.h> /* Required by SunOS4. */
50 #include "file-handle.h"
61 /*#define DEBUGGING 1 */
62 #include "debug-print.h"
64 /* PORTME: This file may require substantial revision for those
65 systems that don't meet the typical 32-bit integer/64-bit double
66 model. It's kinda hard to tell without having one of them on my
69 /* Compression bias used by PSPP. Values between (1 -
70 COMPRESSION_BIAS) and (251 - COMPRESSION_BIAS) inclusive can be
72 #define COMPRESSION_BIAS 100
74 /* sfm writer file_handle extension. */
77 FILE *file; /* Actual file. */
79 int compressed; /* 1=compressed, 0=not compressed. */
80 flt64 *buf; /* Buffered data. */
81 flt64 *end; /* Buffer end. */
82 flt64 *ptr; /* Current location in buffer. */
83 unsigned char *x; /* Location in current instruction octet. */
84 unsigned char *y; /* End of instruction octet. */
85 int n_cases; /* Number of cases written so far. */
87 char *elem_type; /* ALPHA or NUMERIC for each flt64 element. */
90 static struct fh_ext_class sfm_w_class;
92 static char *append_string_max (char *, const char *, const char *);
93 static int write_header (struct sfm_write_info *inf);
94 static int bufwrite (struct file_handle *h, const void *buf, size_t nbytes);
95 static int write_variable (struct sfm_write_info *inf, struct variable *v);
96 static int write_value_labels (struct sfm_write_info *inf, struct variable * s, int index);
97 static int write_rec_7_34 (struct sfm_write_info *inf);
98 static int write_documents (struct sfm_write_info *inf);
100 /* Writes the dictionary INF->dict to system file INF->h. The system
101 file is compressed if INF->compress is nonzero. INF->case_size is
102 set to the number of flt64 elements in a single case. Returns
103 nonzero only if successful. */
105 sfm_write_dictionary (struct sfm_write_info *inf)
107 struct dictionary *d = inf->dict;
108 struct sfm_fhuser_ext *ext;
112 if (inf->h->class != NULL)
114 msg (ME, _("Cannot write file %s as system file: already opened for %s."),
115 fh_handle_name (inf->h), inf->h->class->name);
119 msg (VM (1), _("%s: Opening system-file handle %s for writing."),
120 fh_handle_filename (inf->h), fh_handle_name (inf->h));
122 /* Open the physical disk file. */
123 inf->h->class = &sfm_w_class;
124 inf->h->ext = ext = xmalloc (sizeof (struct sfm_fhuser_ext));
125 ext->file = fopen (inf->h->norm_fn, "wb");
126 ext->elem_type = NULL;
127 if (ext->file == NULL)
129 msg (ME, _("An error occurred while opening \"%s\" for writing "
130 "as a system file: %s."), inf->h->fn, strerror (errno));
136 /* Initialize the sfm_fhuser_ext structure. */
137 ext->compressed = inf->compress;
138 ext->buf = ext->ptr = NULL;
139 ext->x = ext->y = NULL;
142 /* Write the file header. */
143 if (!write_header (inf))
146 /* Write basic variable info. */
147 for (i = 0; i < d->nvar; i++)
148 write_variable (inf, d->var[i]);
150 /* Write out value labels. */
151 for (index = i = 0; i < d->nvar; i++)
153 struct variable *v = d->var[i];
155 if (!write_value_labels (inf, v, index))
157 index += (v->type == NUMERIC ? 1
158 : DIV_RND_UP (v->width, sizeof (flt64)));
161 if (d->documents != NULL && !write_documents (inf))
163 if (!write_rec_7_34 (inf))
166 /* Write record 999. */
175 rec_999.rec_type = 999;
178 if (!bufwrite (inf->h, &rec_999, sizeof rec_999))
182 msg (VM (2), _("Wrote system-file header successfully."));
187 msg (VM (1), _("Error writing system-file header."));
189 inf->h->class = NULL;
191 free (ext->elem_type);
192 ext->elem_type = NULL;
196 /* Returns value of X truncated to two least-significant digits. */
207 /* Write the sysfile_header header to the system file represented by
210 write_header (struct sfm_write_info *inf)
212 struct dictionary *d = inf->dict;
213 struct sfm_fhuser_ext *ext = inf->h->ext;
214 struct sysfile_header hdr;
220 memcpy (hdr.rec_type, "$FL2", 4);
222 p = stpcpy (hdr.prod_name, "@(#) SPSS DATA FILE ");
223 p = append_string_max (p, version, &hdr.prod_name[60]);
224 p = append_string_max (p, " - ", &hdr.prod_name[60]);
225 p = append_string_max (p, host_system, &hdr.prod_name[60]);
226 memset (p, ' ', &hdr.prod_name[60] - p);
231 for (i = 0; i < d->nvar; i++)
233 struct variable *v = d->var[i];
234 hdr.case_size += (v->type == NUMERIC ? 1
235 : DIV_RND_UP (v->width, sizeof (flt64)));
237 inf->case_size = hdr.case_size;
239 p = ext->elem_type = xmalloc (inf->case_size);
240 for (i = 0; i < d->nvar; i++)
242 struct variable *v = d->var[i];
243 int count = (v->type == NUMERIC ? 1
244 : DIV_RND_UP (v->width, sizeof (flt64)));
249 hdr.compressed = inf->compress;
251 update_weighting (d);
252 if (d->weight_index != -1)
254 int recalc_weight_index = 1;
256 for (i = 0; i < d->weight_index; i++)
258 struct variable *v = d->var[i];
259 recalc_weight_index += (v->type == NUMERIC ? 1
260 : DIV_RND_UP (v->width, sizeof (flt64)));
262 hdr.weight_index = recalc_weight_index;
265 hdr.weight_index = 0;
268 hdr.bias = COMPRESSION_BIAS;
270 if ((time_t) - 1 == time (&t))
272 memcpy (hdr.creation_date, "01 Jan 70", 9);
273 memcpy (hdr.creation_time, "00:00:00", 8);
277 static const char *month_name[12] =
279 "Jan", "Feb", "Mar", "Apr", "May", "Jun",
280 "Jul", "Aug", "Sep", "Oct", "Nov", "Dec",
282 struct tm *tmp = localtime (&t);
283 int day = rerange (tmp->tm_mday);
284 int mon = rerange (tmp->tm_mon + 1);
285 int year = rerange (tmp->tm_year);
286 int hour = rerange (tmp->tm_hour + 1);
287 int min = rerange (tmp->tm_min + 1);
288 int sec = rerange (tmp->tm_sec + 1);
291 sprintf (buf, "%02d %s %02d", day, month_name[mon - 1], year);
292 memcpy (hdr.creation_date, buf, sizeof hdr.creation_date);
293 sprintf (buf, "%02d:%02d:%02d", hour - 1, min - 1, sec - 1);
294 memcpy (hdr.creation_time, buf, sizeof hdr.creation_time);
297 st_bare_pad_copy (hdr.file_label, d->label ? d->label : "",
298 sizeof hdr.file_label);
299 memset (hdr.padding, 0, sizeof hdr.padding);
301 if (!bufwrite (inf->h, &hdr, sizeof hdr))
306 /* Translates format spec from internal form in SRC to system file
309 write_format_spec (struct fmt_spec *src, int32 *dest)
311 *dest = (formats[src->type].spss << 16) | (src->w << 8) | src->d;
314 /* Write the variable record(s) for primary variable P and secondary
315 variable S to the system file represented by INF. */
317 write_variable (struct sfm_write_info *inf, struct variable *v)
319 struct sysfile_variable sv;
321 /* Missing values. */
322 flt64 m[3]; /* Missing value values. */
323 int nm; /* Number of missing values, possibly negative. */
326 sv.type = (v->type == NUMERIC ? 0 : v->width);
327 sv.has_var_label = (v->label != NULL);
329 switch (v->miss_type)
337 for (nm = 0; nm < v->miss_type; nm++)
338 m[nm] = v->missing[nm].f;
341 m[0] = v->missing[0].f;
342 m[1] = v->missing[1].f;
346 m[0] = second_lowest_flt64;
347 m[1] = v->missing[0].f;
351 m[0] = v->missing[0].f;
355 case MISSING_RANGE_1:
356 m[0] = v->missing[0].f;
357 m[1] = v->missing[1].f;
358 m[2] = v->missing[2].f;
362 m[0] = second_lowest_flt64;
363 m[1] = v->missing[0].f;
364 m[2] = v->missing[1].f;
368 m[0] = v->missing[0].f;
369 m[1] = second_lowest_flt64;
370 m[2] = v->missing[1].f;
377 sv.n_missing_values = nm;
378 write_format_spec (&v->print, &sv.print);
379 write_format_spec (&v->write, &sv.write);
380 memcpy (sv.name, v->name, strlen (v->name));
381 memset (&sv.name[strlen (v->name)], ' ', 8 - strlen (v->name));
382 if (!bufwrite (inf->h, &sv, sizeof sv))
396 l.label_len = min (strlen (v->label), 120);
397 ext_len = ROUND_UP (l.label_len, sizeof l.label_len);
398 memcpy (l.label, v->label, l.label_len);
399 memset (&l.label[l.label_len], ' ', ext_len - l.label_len);
401 if (!bufwrite (inf->h, &l, offsetof (struct label, label) + ext_len))
405 if (nm && !bufwrite (inf->h, m, sizeof *m * nm))
408 if (v->type == ALPHA && v->width > (int) sizeof (flt64))
414 sv.has_var_label = 0;
415 sv.n_missing_values = 0;
416 memset (&sv.print, 0, sizeof sv.print);
417 memset (&sv.write, 0, sizeof sv.write);
418 memset (&sv.name, 0, sizeof sv.name);
420 pad_count = DIV_RND_UP (v->width, (int) sizeof (flt64)) - 1;
421 for (i = 0; i < pad_count; i++)
422 if (!bufwrite (inf->h, &sv, sizeof sv))
429 /* Writes the value labels for variable V having system file variable
430 index INDEX to the system file associated with INF. Returns
431 nonzero only if successful. */
433 write_value_labels (struct sfm_write_info * inf, struct variable *v, int index)
435 struct value_label_rec
442 struct variable_index_rec
450 struct value_label_rec *vlr;
451 struct variable_index_rec vir;
452 struct value_label *vl;
455 avl_traverser_init (i);
457 if (v->val_lab == NULL || avl_count (v->val_lab) == 0)
460 /* Pass 1: Count bytes. */
461 vlr_size = (sizeof (struct value_label_rec)
462 + sizeof (flt64) * (avl_count (v->val_lab) - 1));
463 while (NULL != (vl = avl_traverse (v->val_lab, &i)))
464 vlr_size += ROUND_UP (strlen (vl->s) + 1, sizeof (flt64));
466 /* Pass 2: Copy bytes. */
467 vlr = local_alloc (vlr_size);
469 vlr->n_labels = avl_count (v->val_lab);
471 while (NULL != (vl = avl_traverse (v->val_lab, &i)))
473 int len = strlen (vl->s);
476 *(unsigned char *) loc = len;
477 memcpy (&((unsigned char *) loc)[1], vl->s, len);
478 memset (&((unsigned char *) loc)[1 + len], ' ',
479 REM_RND_UP (len + 1, sizeof (flt64)));
480 loc += DIV_RND_UP (len + 1, sizeof (flt64));
483 if (!bufwrite (inf->h, vlr, vlr_size))
492 vir.vars[0] = index + 1;
493 if (!bufwrite (inf->h, &vir, sizeof vir))
499 /* Writes record type 6, document record. */
501 write_documents (struct sfm_write_info * inf)
503 struct dictionary *d = inf->dict;
506 int32 rec_type P; /* Always 6. */
507 int32 n_lines P; /* Number of lines of documents. */
512 rec_6.n_lines = d->n_documents;
513 if (!bufwrite (inf->h, &rec_6, sizeof rec_6))
515 if (!bufwrite (inf->h, d->documents, 80 * d->n_documents))
521 /* Writes record type 7, subtypes 3 and 4. */
523 write_rec_7_34 (struct sfm_write_info * inf)
540 /* Components of the version number, from major to minor. */
541 int version_component[3];
543 /* Used to step through the version string. */
546 /* Parses the version string, which is assumed to be of the form
547 #.#x, where each # is a string of digits, and x is a single
549 version_component[0] = strtol (bare_version, &p, 10);
552 version_component[1] = strtol (bare_version, &p, 10);
553 version_component[2] = (isalpha ((unsigned char) *p)
554 ? tolower ((unsigned char) *p) - 'a' : 0);
556 rec_7.rec_type_3 = 7;
558 rec_7.data_type_3 = sizeof (int32);
560 rec_7.elem_3[0] = version_component[0];
561 rec_7.elem_3[1] = version_component[1];
562 rec_7.elem_3[2] = version_component[2];
563 rec_7.elem_3[3] = -1;
565 /* PORTME: 1=IEEE754, 2=IBM 370, 3=DEC VAX E. */
566 #if FPREP==FPREP_IEEE754
569 #error Floating-point representation unknown here.
574 /* PORTME: 1=big-endian, 2=little-endian. */
577 else if (endian == LITTLE)
582 /* PORTME: 1=EBCDIC, 2=7-bit ASCII, 3=8-bit ASCII, 4=DEC Kanji. */
585 rec_7.rec_type_4 = 7;
587 rec_7.data_type_4 = sizeof (flt64);
589 rec_7.elem_4[0] = -FLT64_MAX;
590 rec_7.elem_4[1] = FLT64_MAX;
591 rec_7.elem_4[2] = second_lowest_flt64;
593 if (!bufwrite (inf->h, &rec_7, sizeof rec_7))
598 /* Write NBYTES starting at BUF to the system file represented by
601 bufwrite (struct file_handle * h, const void *buf, size_t nbytes)
603 struct sfm_fhuser_ext *ext = h->ext;
606 if (1 != fwrite (buf, nbytes, 1, ext->file))
608 msg (ME, _("%s: Writing system file: %s."), h->fn, strerror (errno));
614 /* Copies string DEST to SRC with the proviso that DEST does not reach
615 byte END; no null terminator is copied. Returns a pointer to the
616 byte after the last byte copied. */
618 append_string_max (char *dest, const char *src, const char *end)
620 int nbytes = min (end - dest, (int) strlen (src));
621 memcpy (dest, src, nbytes);
622 return dest + nbytes;
625 /* Makes certain that the compression buffer of H has room for another
626 element. If there's not room, pads out the current instruction
627 octet with zero and dumps out the buffer. */
629 ensure_buf_space (struct file_handle *h)
631 struct sfm_fhuser_ext *ext = h->ext;
633 if (ext->ptr >= ext->end)
635 memset (ext->x, 0, ext->y - ext->x);
638 if (!bufwrite (h, ext->buf, sizeof *ext->buf * 128))
644 /* Writes case ELEM consisting of N_ELEM flt64 elements to the system
645 file represented by H. Return success. */
647 sfm_write_case (struct file_handle * h, const flt64 *elem, int n_elem)
649 struct sfm_fhuser_ext *ext = h->ext;
650 const flt64 *end_elem = &elem[n_elem];
651 char *elem_type = ext->elem_type;
655 if (ext->compressed == 0)
656 return bufwrite (h, elem, sizeof *elem * n_elem);
658 if (ext->buf == NULL)
660 ext->buf = xmalloc (sizeof *ext->buf * 128);
662 ext->end = &ext->buf[128];
663 ext->x = (unsigned char *) (ext->ptr++);
664 ext->y = (unsigned char *) (ext->ptr);
666 for (; elem < end_elem; elem++, elem_type++)
668 if (ext->x >= ext->y)
670 if (!ensure_buf_space (h))
672 ext->x = (unsigned char *) (ext->ptr++);
673 ext->y = (unsigned char *) (ext->ptr);
676 if (*elem_type == NUMERIC)
678 if (*elem == -FLT64_MAX)
685 int value = *elem < 0 ? *elem - EPSILON : *elem + EPSILON;
687 if (value >= 1 - COMPRESSION_BIAS
688 && value <= 251 - COMPRESSION_BIAS
689 && approx_eq (value, *elem))
691 *ext->x++ = value + COMPRESSION_BIAS;
698 if (0 == memcmp ((char *) elem,
708 if (!ensure_buf_space (h))
716 /* Closes a system file after we're done with it. */
718 sfm_close (struct file_handle * h)
720 struct sfm_fhuser_ext *ext = h->ext;
722 if (ext->buf != NULL && ext->ptr > ext->buf)
724 memset (ext->x, 0, ext->y - ext->x);
725 bufwrite (h, ext->buf, (ext->ptr - ext->buf) * sizeof *ext->buf);
728 /* Attempt to seek back to the beginning in order to write the
729 number of cases. If that's not possible (i.e., we're writing to
730 a tty or a pipe), then it's not a big deal because we wrote the
731 code that indicates an unknown number of cases. */
732 if (0 == fseek (ext->file, offsetof (struct sysfile_header, ncases),
735 int32 n_cases = ext->n_cases;
737 /* I don't really care about the return value: it doesn't matter
738 whether this data is written. This is the only situation in
739 which you will see me fail to check a return value. */
740 fwrite (&n_cases, sizeof n_cases, 1, ext->file);
743 if (EOF == fclose (ext->file))
744 msg (ME, _("%s: Closing system file: %s."), h->fn, strerror (errno));
747 free (ext->elem_type);
751 static struct fh_ext_class sfm_w_class =
754 N_("writing as a system file"),