1 /* PSPP - computes sample statistics.
2 Copyright (C) 1997-9, 2000, 2006 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., 51 Franklin Street, Fifth Floor, Boston, MA
27 #include <data/data-in.h>
28 #include <data/data-out.h>
29 #include <data/dictionary.h>
30 #include <data/format.h>
31 #include <data/procedure.h>
32 #include <data/settings.h>
33 #include <data/variable.h>
34 #include <language/command.h>
35 #include <language/lexer/format-parser.h>
36 #include <language/lexer/lexer.h>
37 #include <language/prompt.h>
38 #include <libpspp/alloc.h>
39 #include <libpspp/compiler.h>
40 #include <libpspp/copyleft.h>
41 #include <libpspp/float-format.h>
42 #include <libpspp/integer-format.h>
43 #include <libpspp/magic.h>
44 #include <libpspp/message.h>
45 #include <math/random.h>
46 #include <output/output.h>
51 #else /* !HAVE_TERMCAP_H */
52 int tgetent (char *, const char *);
53 int tgetnum (const char *);
54 #endif /* !HAVE_TERMCAP_H */
55 #endif /* !HAVE_LIBTERMCAP */
58 #define _(msgid) gettext (msgid)
63 block=string "x==1" "one character long";
64 boxstring=string "x==3 || x==11" "3 or 11 characters long";
65 case=size:upper/uplow;
71 compression=compress:on/off;
72 cpi=integer "x>0" "%s must be greater than 0";
74 decimal=dec:dot/comma;
78 endcmd=string "x==1" "one character long";
80 errorbreak=errbrk:on/off;
81 errors=errors:terminal/listing/both/on/none/off;
83 headers=headers:no/yes/blank;
85 histogram=string "x==1" "one character long";
90 lowres=lores:auto/on/off;
91 lpi=integer "x>0" "%s must be greater than 0";
92 menus=menus:standard/extended;
93 messages=messages:on/off/terminal/listing/both/on/none/off;
95 miterate=integer "x>0" "%s must be greater than 0";
96 mnest=integer "x>0" "%s must be greater than 0";
98 mxerrs=integer "x >= 1" "%s must be at least 1";
99 mxloops=integer "x >=1" "%s must be at least 1";
103 printback=prtbck:on/off;
105 results=res:on/off/terminal/listing/both/on/none/off;
106 rib=rib:msbfirst/lsbfirst/vax/native;
107 rrb=rrb:native/isl/isb/idl/idb/vf/vd/vg/zs/zl;
109 scompression=scompress:on/off;
110 scripttab=string "x==1" "one character long";
112 tb1=string "x==3 || x==11" "3 or 11 characters long";
114 undefined=undef:warn/nowarn;
115 wib=wib:msbfirst/lsbfirst/vax/native;
116 wrb=wrb:native/isl/isb/idl/idb/vf/vd/vg/zs/zl;
118 workspace=integer "x>=1024" "%s must be at least 1 MB";
128 static bool do_cc (const char *cc_string, enum fmt_type);
129 static enum integer_format stc_to_integer_format (int stc);
130 static enum float_format stc_to_float_format (int stc);
133 cmd_set (struct lexer *lexer, struct dataset *ds)
137 if (!parse_set (lexer, ds, &cmd, NULL))
144 do_cc (cmd.s_cca, FMT_CCA);
146 do_cc (cmd.s_ccb, FMT_CCB);
148 do_cc (cmd.s_ccc, FMT_CCC);
150 do_cc (cmd.s_ccd, FMT_CCD);
152 do_cc (cmd.s_cce, FMT_CCE);
155 prompt_set (PROMPT_FIRST, cmd.s_prompt);
157 prompt_set (PROMPT_LATER, cmd.s_cprompt);
159 prompt_set (PROMPT_DATA, cmd.s_dprompt);
162 fmt_set_decimal (cmd.dec == STC_DOT ? '.' : ',');
164 set_echo (cmd.echo == STC_ON);
166 set_endcmd (cmd.s_endcmd[0]);
167 if (cmd.sbc_errorbreak)
168 set_errorbreak (cmd.errbrk == STC_ON);
171 bool both = cmd.errors == STC_BOTH || cmd.errors == STC_ON;
172 set_error_routing_to_terminal (cmd.errors == STC_TERMINAL || both);
173 set_error_routing_to_listing (cmd.errors == STC_LISTING || both);
176 set_include (cmd.inc == STC_ON);
178 set_mxerrs (cmd.n_mxerrs[0]);
180 set_mxwarns (cmd.n_mxwarns[0]);
182 set_nulline (cmd.null == STC_ON);
184 data_in_set_integer_format (stc_to_integer_format (cmd.rib));
186 data_in_set_float_format (stc_to_float_format (cmd.rrb));
189 if (cmd.sbc_scompression)
190 set_scompression (cmd.scompress == STC_ON);
191 if (cmd.sbc_undefined)
192 set_undefined (cmd.undef == STC_WARN);
194 data_out_set_integer_format (stc_to_integer_format (cmd.wib));
196 data_out_set_float_format (stc_to_float_format (cmd.wrb));
197 if (cmd.sbc_workspace)
198 set_workspace (cmd.n_workspace[0] * 1024L);
201 msg (SW, _("%s is obsolete."), "BLOCK");
202 if (cmd.sbc_boxstring)
203 msg (SW, _("%s is obsolete."), "BOXSTRING");
204 if (cmd.sbc_histogram)
205 msg (SW, _("%s is obsolete."), "HISTOGRAM");
207 msg (SW, _("%s is obsolete."), "MENUS");
209 msg (SW, _("%s is obsolete."), "XSORT");
210 if (cmd.sbc_mxmemory)
211 msg (SE, _("%s is obsolete."), "MXMEMORY");
212 if (cmd.sbc_scripttab)
213 msg (SE, _("%s is obsolete."), "SCRIPTTAB");
215 msg (SW, _("%s is obsolete."), "TBFONTS");
216 if (cmd.sbc_tb1 && cmd.s_tb1)
217 msg (SW, _("%s is obsolete."), "TB1");
220 msg (SW, _("%s is not implemented."), "CASE");
222 if (cmd.sbc_compression)
223 msg (SW, _("Active file compression is not implemented."));
230 /* Returns the integer_format value corresponding to STC,
231 which should be the value of cmd.rib or cmd.wib. */
232 static enum integer_format
233 stc_to_integer_format (int stc)
235 return (stc == STC_MSBFIRST ? INTEGER_MSB_FIRST
236 : stc == STC_LSBFIRST ? INTEGER_LSB_FIRST
237 : stc == STC_VAX ? INTEGER_VAX
241 /* Returns the float_format value corresponding to STC,
242 which should be the value of cmd.rrb or cmd.wrb. */
243 static enum float_format
244 stc_to_float_format (int stc)
249 return FLOAT_NATIVE_DOUBLE;
252 return FLOAT_IEEE_SINGLE_LE;
254 return FLOAT_IEEE_SINGLE_BE;
256 return FLOAT_IEEE_DOUBLE_LE;
258 return FLOAT_IEEE_DOUBLE_BE;
268 return FLOAT_Z_SHORT;
276 /* Find the grouping characters in CC_STRING and set CC's
277 grouping and decimal members appropriately. Returns true if
278 successful, false otherwise. */
280 find_cc_separators (const char *cc_string, struct fmt_number_style *cc)
283 int comma_cnt, dot_cnt;
285 /* Count commas and periods. There must be exactly three of
286 one or the other, except that an apostrophe escapes a
287 following comma or period. */
288 comma_cnt = dot_cnt = 0;
289 for (sp = cc_string; *sp; sp++)
294 else if (*sp == '\'' && (sp[1] == '.' || sp[1] == ',' || sp[1] == '\''))
297 if ((comma_cnt == 3) == (dot_cnt == 3))
313 /* Extracts a token from IN into a newly allocated AFFIX. Tokens
314 are delimited by GROUPING. The token is truncated to at most
315 FMT_STYLE_AFFIX_MAX characters. Returns the first character
316 following the token. */
318 extract_cc_token (const char *in, int grouping, struct substring *affix)
321 ss_alloc_uninit (affix, FMT_STYLE_AFFIX_MAX);
322 for (; *in != '\0' && *in != grouping; in++)
324 if (*in == '\'' && in[1] == grouping)
326 if (ofs < FMT_STYLE_AFFIX_MAX)
327 ss_data (*affix)[ofs++] = *in;
336 /* Sets custom currency specifier CC having name CC_NAME ('A' through
337 'E') to correspond to the settings in CC_STRING. */
339 do_cc (const char *cc_string, enum fmt_type type)
341 struct fmt_number_style *cc = fmt_number_style_create ();
343 /* Determine separators. */
344 if (!find_cc_separators (cc_string, cc))
346 fmt_number_style_destroy (cc);
347 msg (SE, _("%s: Custom currency string `%s' does not contain "
348 "exactly three periods or commas (or it contains both)."),
349 fmt_name (type), cc_string);
353 cc_string = extract_cc_token (cc_string, cc->grouping, &cc->neg_prefix);
354 cc_string = extract_cc_token (cc_string, cc->grouping, &cc->prefix);
355 cc_string = extract_cc_token (cc_string, cc->grouping, &cc->suffix);
356 cc_string = extract_cc_token (cc_string, cc->grouping, &cc->neg_suffix);
358 fmt_set_style (type, cc);
363 /* Parses the BLANKS subcommand, which controls the value that
364 completely blank fields in numeric data imply. X, Wnd: Syntax is
365 SYSMIS or a numeric value. */
367 stc_custom_blanks (struct lexer *lexer,
368 struct dataset *ds UNUSED,
369 struct cmd_set *cmd UNUSED, void *aux UNUSED)
371 lex_match (lexer, '=');
372 if (lex_match_id (lexer, "SYSMIS"))
379 if (!lex_force_num (lexer))
381 set_blanks (lex_number (lexer));
387 /* Parses the EPOCH subcommand, which controls the epoch used for
388 parsing 2-digit years. */
390 stc_custom_epoch (struct lexer *lexer,
391 struct dataset *ds UNUSED,
392 struct cmd_set *cmd UNUSED, void *aux UNUSED)
394 lex_match (lexer, '=');
395 if (lex_match_id (lexer, "AUTOMATIC"))
397 else if (lex_is_integer (lexer))
399 int new_epoch = lex_integer (lexer);
401 if (new_epoch < 1500)
403 msg (SE, _("EPOCH must be 1500 or later."));
406 set_epoch (new_epoch);
410 lex_error (lexer, _("expecting AUTOMATIC or year"));
418 stc_custom_length (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
422 lex_match (lexer, '=');
423 if (lex_match_id (lexer, "NONE"))
427 if (!lex_force_int (lexer))
429 if (lex_integer (lexer) < 1)
431 msg (SE, _("LENGTH must be at least 1."));
434 page_length = lex_integer (lexer);
438 if (page_length != -1)
439 set_viewlength (page_length);
445 stc_custom_seed (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
447 lex_match (lexer, '=');
448 if (lex_match_id (lexer, "RANDOM"))
452 if (!lex_force_num (lexer))
454 set_rng (lex_number (lexer));
462 stc_custom_width (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
464 lex_match (lexer, '=');
465 if (lex_match_id (lexer, "NARROW"))
467 else if (lex_match_id (lexer, "WIDE"))
471 if (!lex_force_int (lexer))
473 if (lex_integer (lexer) < 40)
475 msg (SE, _("WIDTH must be at least 40."));
478 set_viewwidth (lex_integer (lexer));
485 /* Parses FORMAT subcommand, which consists of a numeric format
488 stc_custom_format (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
492 lex_match (lexer, '=');
493 if (!parse_format_specifier (lexer, &fmt))
495 if (fmt_is_string (fmt.type))
497 char str[FMT_STRING_LEN_MAX + 1];
498 msg (SE, _("FORMAT requires numeric output format as an argument. "
499 "Specified format %s is of type string."),
500 fmt_to_string (&fmt, str));
509 stc_custom_journal (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
511 lex_match (lexer, '=');
512 if (!lex_match_id (lexer, "ON") && !lex_match_id (lexer, "OFF"))
514 if (lex_token (lexer) == T_STRING)
518 lex_error (lexer, NULL);
526 stc_custom_listing (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
530 lex_match (lexer, '=');
531 if (lex_match_id (lexer, "ON") || lex_match_id (lexer, "YES"))
533 else if (lex_match_id (lexer, "OFF") || lex_match_id (lexer, "NO"))
540 outp_enable_device (listing, OUTP_DEV_LISTING);
546 stc_custom_disk (struct lexer *lexer, struct dataset *ds, struct cmd_set *cmd UNUSED, void *aux)
548 return stc_custom_listing (lexer, ds, cmd, aux);
552 show_blanks (const struct dataset *ds UNUSED)
554 if (get_blanks () == SYSMIS)
555 msg (SN, _("BLANKS is SYSMIS."));
557 msg (SN, _("BLANKS is %g."), get_blanks ());
562 format_cc (struct substring in, char grouping, char *out)
564 while (!ss_is_empty (in))
566 char c = ss_get_char (&in);
567 if (c == grouping || c == '\'')
577 show_cc (enum fmt_type type)
579 const struct fmt_number_style *cc = fmt_get_style (type);
580 char cc_string[FMT_STYLE_AFFIX_MAX * 4 * 2 + 3 + 1];
583 out = format_cc (cc->neg_prefix, cc->grouping, cc_string);
584 *out++ = cc->grouping;
585 out = format_cc (cc->prefix, cc->grouping, out);
586 *out++ = cc->grouping;
587 out = format_cc (cc->suffix, cc->grouping, out);
588 *out++ = cc->grouping;
589 out = format_cc (cc->neg_suffix, cc->grouping, out);
592 msg (SN, _("%s is \"%s\"."), fmt_name (type), cc_string);
596 show_cca (const struct dataset *ds UNUSED)
602 show_ccb (const struct dataset *ds UNUSED)
608 show_ccc (const struct dataset *ds UNUSED)
614 show_ccd (const struct dataset *ds UNUSED)
620 show_cce (const struct dataset *ds UNUSED)
626 show_decimals (const struct dataset *ds UNUSED)
628 msg (SN, _("DECIMAL is \"%c\"."), fmt_decimal_char (FMT_F));
632 show_endcmd (const struct dataset *ds UNUSED)
634 msg (SN, _("ENDCMD is \"%c\"."), get_endcmd ());
638 show_errors (const struct dataset *ds UNUSED)
640 bool terminal = get_error_routing_to_terminal ();
641 bool listing = get_error_routing_to_listing ();
642 msg (SN, _("ERRORS is \"%s\"."),
643 terminal && listing ? "BOTH"
644 : terminal ? "TERMINAL"
645 : listing ? "LISTING"
650 show_format (const struct dataset *ds UNUSED)
652 char str[FMT_STRING_LEN_MAX + 1];
653 msg (SN, _("FORMAT is %s."), fmt_to_string (get_format (), str));
657 show_length (const struct dataset *ds UNUSED)
659 msg (SN, _("LENGTH is %d."), get_viewlength ());
663 show_mxerrs (const struct dataset *ds UNUSED)
665 msg (SN, _("MXERRS is %d."), get_mxerrs ());
669 show_mxloops (const struct dataset *ds UNUSED)
671 msg (SN, _("MXLOOPS is %d."), get_mxloops ());
675 show_mxwarns (const struct dataset *ds UNUSED)
677 msg (SN, _("MXWARNS is %d."), get_mxwarns ());
680 /* Outputs that SETTING has the given INTEGER_FORMAT value. */
682 show_integer_format (const char *setting, enum integer_format integer_format)
684 msg (SN, _("%s is %s (%s)."),
686 (integer_format == INTEGER_MSB_FIRST ? "MSBFIRST"
687 : integer_format == INTEGER_LSB_FIRST ? "LSBFIRST"
689 integer_format == INTEGER_NATIVE ? "NATIVE" : "nonnative");
692 /* Outputs that SETTING has the given FLOAT_FORMAT value. */
694 show_float_format (const char *setting, enum float_format float_format)
696 const char *format_name = "";
698 switch (float_format)
700 case FLOAT_IEEE_SINGLE_LE:
701 format_name = "ISL (32-bit IEEE 754 single, little-endian)";
703 case FLOAT_IEEE_SINGLE_BE:
704 format_name = "ISB (32-bit IEEE 754 single, big-endian)";
706 case FLOAT_IEEE_DOUBLE_LE:
707 format_name = "IDL (64-bit IEEE 754 double, little-endian)";
709 case FLOAT_IEEE_DOUBLE_BE:
710 format_name = "IDB (64-bit IEEE 754 double, big-endian)";
714 format_name = "VF (32-bit VAX F, VAX-endian)";
717 format_name = "VD (64-bit VAX D, VAX-endian)";
720 format_name = "VG (64-bit VAX G, VAX-endian)";
724 format_name = "ZS (32-bit IBM Z hexadecimal short, big-endian)";
727 format_name = "ZL (64-bit IBM Z hexadecimal long, big-endian)";
735 msg (SN, _("%s is %s (%s)."),
736 setting, format_name,
737 float_format == FLOAT_NATIVE_DOUBLE ? "NATIVE" : "nonnative");
741 show_rib (const struct dataset *ds UNUSED)
743 show_integer_format ("RIB", data_in_get_integer_format ());
747 show_rrb (const struct dataset *ds UNUSED)
749 show_float_format ("RRB", data_in_get_float_format ());
753 show_scompression (const struct dataset *ds UNUSED)
755 if (get_scompression ())
756 msg (SN, _("SCOMPRESSION is ON."));
758 msg (SN, _("SCOMPRESSION is OFF."));
762 show_undefined (const struct dataset *ds UNUSED)
764 if (get_undefined ())
765 msg (SN, _("UNDEFINED is WARN."));
767 msg (SN, _("UNDEFINED is NOWARN."));
771 show_weight (const struct dataset *ds)
773 struct variable *var = dict_get_weight (dataset_dict (ds));
775 msg (SN, _("WEIGHT is off."));
777 msg (SN, _("WEIGHT is variable %s."), var_get_name (var));
781 show_wib (const struct dataset *ds UNUSED)
783 show_integer_format ("WIB", data_out_get_integer_format ());
787 show_wrb (const struct dataset *ds UNUSED)
789 show_float_format ("WRB", data_out_get_float_format ());
793 show_width (const struct dataset *ds UNUSED)
795 msg (SN, _("WIDTH is %d."), get_viewwidth ());
801 void (*function) (const struct dataset *);
804 const struct show_sbc show_table[] =
806 {"BLANKS", show_blanks},
812 {"DECIMALS", show_decimals},
813 {"ENDCMD", show_endcmd},
814 {"ERRORS", show_errors},
815 {"FORMAT", show_format},
816 {"LENGTH", show_length},
817 {"MXERRS", show_mxerrs},
818 {"MXLOOPS", show_mxloops},
819 {"MXWARNS", show_mxwarns},
822 {"SCOMPRESSION", show_scompression},
823 {"UNDEFINED", show_undefined},
824 {"WEIGHT", show_weight},
827 {"WIDTH", show_width},
831 show_all (const struct dataset *ds)
835 for (i = 0; i < sizeof show_table / sizeof *show_table; i++)
836 show_table[i].function (ds);
844 for (i = 0; i < 5; i++)
849 show_warranty (const struct dataset *ds UNUSED)
851 msg (MN, lack_of_warranty);
855 show_copying (const struct dataset *ds UNUSED)
861 cmd_show (struct lexer *lexer, struct dataset *ds)
863 if (lex_token (lexer) == '.')
871 if (lex_match (lexer, T_ALL))
873 else if (lex_match_id (lexer, "CC"))
875 else if (lex_match_id (lexer, "WARRANTY"))
877 else if (lex_match_id (lexer, "COPYING"))
879 else if (lex_token (lexer) == T_ID)
883 for (i = 0; i < sizeof show_table / sizeof *show_table; i++)
884 if (lex_match_id (lexer, show_table[i].name))
886 show_table[i].function (ds);
889 lex_error (lexer, NULL);
896 lex_error (lexer, NULL);
900 lex_match (lexer, '/');
902 while (lex_token (lexer) != '.');