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/value.h>
34 #include <data/variable.h>
35 #include <language/command.h>
36 #include <language/lexer/format-parser.h>
37 #include <language/lexer/lexer.h>
38 #include <language/prompt.h>
39 #include <libpspp/alloc.h>
40 #include <libpspp/compiler.h>
41 #include <libpspp/copyleft.h>
42 #include <libpspp/float-format.h>
43 #include <libpspp/integer-format.h>
44 #include <libpspp/magic.h>
45 #include <libpspp/message.h>
46 #include <math/random.h>
47 #include <output/output.h>
52 #else /* !HAVE_TERMCAP_H */
53 int tgetent (char *, const char *);
54 int tgetnum (const char *);
55 #endif /* !HAVE_TERMCAP_H */
56 #endif /* !HAVE_LIBTERMCAP */
59 #define _(msgid) gettext (msgid)
64 block=string "x==1" "one character long";
65 boxstring=string "x==3 || x==11" "3 or 11 characters long";
66 case=size:upper/uplow;
72 compression=compress:on/off;
73 cpi=integer "x>0" "%s must be greater than 0";
75 decimal=dec:dot/comma;
79 endcmd=string "x==1" "one character long";
81 errorbreak=errbrk:on/off;
82 errors=errors:terminal/listing/both/on/none/off;
84 headers=headers:no/yes/blank;
86 histogram=string "x==1" "one character long";
91 lowres=lores:auto/on/off;
92 lpi=integer "x>0" "%s must be greater than 0";
93 menus=menus:standard/extended;
94 messages=messages:on/off/terminal/listing/both/on/none/off;
96 miterate=integer "x>0" "%s must be greater than 0";
97 mnest=integer "x>0" "%s must be greater than 0";
99 mxerrs=integer "x >= 1" "%s must be at least 1";
100 mxloops=integer "x >=1" "%s must be at least 1";
104 printback=prtbck:on/off;
106 results=res:on/off/terminal/listing/both/on/none/off;
107 rib=rib:msbfirst/lsbfirst/vax/native;
108 rrb=rrb:native/isl/isb/idl/idb/vf/vd/vg/zs/zl;
110 scompression=scompress:on/off;
111 scripttab=string "x==1" "one character long";
113 tb1=string "x==3 || x==11" "3 or 11 characters long";
115 undefined=undef:warn/nowarn;
116 wib=wib:msbfirst/lsbfirst/vax/native;
117 wrb=wrb:native/isl/isb/idl/idb/vf/vd/vg/zs/zl;
119 workspace=integer "x>=1024" "%s must be at least 1 MB";
129 static bool do_cc (const char *cc_string, enum fmt_type);
130 static enum integer_format stc_to_integer_format (int stc);
131 static enum float_format stc_to_float_format (int stc);
134 cmd_set (struct lexer *lexer, struct dataset *ds)
138 if (!parse_set (lexer, ds, &cmd, NULL))
145 do_cc (cmd.s_cca, FMT_CCA);
147 do_cc (cmd.s_ccb, FMT_CCB);
149 do_cc (cmd.s_ccc, FMT_CCC);
151 do_cc (cmd.s_ccd, FMT_CCD);
153 do_cc (cmd.s_cce, FMT_CCE);
156 prompt_set (PROMPT_FIRST, cmd.s_prompt);
158 prompt_set (PROMPT_LATER, cmd.s_cprompt);
160 prompt_set (PROMPT_DATA, cmd.s_dprompt);
163 fmt_set_decimal (cmd.dec == STC_DOT ? '.' : ',');
165 set_echo (cmd.echo == STC_ON);
167 set_endcmd (cmd.s_endcmd[0]);
168 if (cmd.sbc_errorbreak)
169 set_errorbreak (cmd.errbrk == STC_ON);
172 bool both = cmd.errors == STC_BOTH || cmd.errors == STC_ON;
173 set_error_routing_to_terminal (cmd.errors == STC_TERMINAL || both);
174 set_error_routing_to_listing (cmd.errors == STC_LISTING || both);
177 set_include (cmd.inc == STC_ON);
179 set_mxerrs (cmd.n_mxerrs[0]);
181 set_mxwarns (cmd.n_mxwarns[0]);
183 set_nulline (cmd.null == STC_ON);
185 data_in_set_integer_format (stc_to_integer_format (cmd.rib));
187 data_in_set_float_format (stc_to_float_format (cmd.rrb));
190 if (cmd.sbc_scompression)
191 set_scompression (cmd.scompress == STC_ON);
192 if (cmd.sbc_undefined)
193 set_undefined (cmd.undef == STC_WARN);
195 data_out_set_integer_format (stc_to_integer_format (cmd.wib));
197 data_out_set_float_format (stc_to_float_format (cmd.wrb));
198 if (cmd.sbc_workspace)
199 set_workspace (cmd.n_workspace[0] * 1024L);
202 msg (SW, _("%s is obsolete."), "BLOCK");
203 if (cmd.sbc_boxstring)
204 msg (SW, _("%s is obsolete."), "BOXSTRING");
205 if (cmd.sbc_histogram)
206 msg (SW, _("%s is obsolete."), "HISTOGRAM");
208 msg (SW, _("%s is obsolete."), "MENUS");
210 msg (SW, _("%s is obsolete."), "XSORT");
211 if (cmd.sbc_mxmemory)
212 msg (SE, _("%s is obsolete."), "MXMEMORY");
213 if (cmd.sbc_scripttab)
214 msg (SE, _("%s is obsolete."), "SCRIPTTAB");
216 msg (SW, _("%s is obsolete."), "TBFONTS");
217 if (cmd.sbc_tb1 && cmd.s_tb1)
218 msg (SW, _("%s is obsolete."), "TB1");
221 msg (SW, _("%s is not implemented."), "CASE");
223 if (cmd.sbc_compression)
224 msg (SW, _("Active file compression is not implemented."));
231 /* Returns the integer_format value corresponding to STC,
232 which should be the value of cmd.rib or cmd.wib. */
233 static enum integer_format
234 stc_to_integer_format (int stc)
236 return (stc == STC_MSBFIRST ? INTEGER_MSB_FIRST
237 : stc == STC_LSBFIRST ? INTEGER_LSB_FIRST
238 : stc == STC_VAX ? INTEGER_VAX
242 /* Returns the float_format value corresponding to STC,
243 which should be the value of cmd.rrb or cmd.wrb. */
244 static enum float_format
245 stc_to_float_format (int stc)
250 return FLOAT_NATIVE_DOUBLE;
253 return FLOAT_IEEE_SINGLE_LE;
255 return FLOAT_IEEE_SINGLE_BE;
257 return FLOAT_IEEE_DOUBLE_LE;
259 return FLOAT_IEEE_DOUBLE_BE;
269 return FLOAT_Z_SHORT;
277 /* Find the grouping characters in CC_STRING and set CC's
278 grouping and decimal members appropriately. Returns true if
279 successful, false otherwise. */
281 find_cc_separators (const char *cc_string, struct fmt_number_style *cc)
284 int comma_cnt, dot_cnt;
286 /* Count commas and periods. There must be exactly three of
287 one or the other, except that an apostrophe escapes a
288 following comma or period. */
289 comma_cnt = dot_cnt = 0;
290 for (sp = cc_string; *sp; sp++)
295 else if (*sp == '\'' && (sp[1] == '.' || sp[1] == ',' || sp[1] == '\''))
298 if ((comma_cnt == 3) == (dot_cnt == 3))
314 /* Extracts a token from IN into a newly allocated AFFIX. Tokens
315 are delimited by GROUPING. The token is truncated to at most
316 FMT_STYLE_AFFIX_MAX characters. Returns the first character
317 following the token. */
319 extract_cc_token (const char *in, int grouping, struct substring *affix)
322 ss_alloc_uninit (affix, FMT_STYLE_AFFIX_MAX);
323 for (; *in != '\0' && *in != grouping; in++)
325 if (*in == '\'' && in[1] == grouping)
327 if (ofs < FMT_STYLE_AFFIX_MAX)
328 ss_data (*affix)[ofs++] = *in;
337 /* Sets custom currency specifier CC having name CC_NAME ('A' through
338 'E') to correspond to the settings in CC_STRING. */
340 do_cc (const char *cc_string, enum fmt_type type)
342 struct fmt_number_style *cc = fmt_number_style_create ();
344 /* Determine separators. */
345 if (!find_cc_separators (cc_string, cc))
347 fmt_number_style_destroy (cc);
348 msg (SE, _("%s: Custom currency string `%s' does not contain "
349 "exactly three periods or commas (or it contains both)."),
350 fmt_name (type), cc_string);
354 cc_string = extract_cc_token (cc_string, cc->grouping, &cc->neg_prefix);
355 cc_string = extract_cc_token (cc_string, cc->grouping, &cc->prefix);
356 cc_string = extract_cc_token (cc_string, cc->grouping, &cc->suffix);
357 cc_string = extract_cc_token (cc_string, cc->grouping, &cc->neg_suffix);
359 fmt_set_style (type, cc);
364 /* Parses the BLANKS subcommand, which controls the value that
365 completely blank fields in numeric data imply. X, Wnd: Syntax is
366 SYSMIS or a numeric value. */
368 stc_custom_blanks (struct lexer *lexer,
369 struct dataset *ds UNUSED,
370 struct cmd_set *cmd UNUSED, void *aux UNUSED)
372 lex_match (lexer, '=');
373 if (lex_match_id (lexer, "SYSMIS"))
380 if (!lex_force_num (lexer))
382 set_blanks (lex_number (lexer));
388 /* Parses the EPOCH subcommand, which controls the epoch used for
389 parsing 2-digit years. */
391 stc_custom_epoch (struct lexer *lexer,
392 struct dataset *ds UNUSED,
393 struct cmd_set *cmd UNUSED, void *aux UNUSED)
395 lex_match (lexer, '=');
396 if (lex_match_id (lexer, "AUTOMATIC"))
398 else if (lex_is_integer (lexer))
400 int new_epoch = lex_integer (lexer);
402 if (new_epoch < 1500)
404 msg (SE, _("EPOCH must be 1500 or later."));
407 set_epoch (new_epoch);
411 lex_error (lexer, _("expecting AUTOMATIC or year"));
419 stc_custom_length (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
423 lex_match (lexer, '=');
424 if (lex_match_id (lexer, "NONE"))
428 if (!lex_force_int (lexer))
430 if (lex_integer (lexer) < 1)
432 msg (SE, _("LENGTH must be at least 1."));
435 page_length = lex_integer (lexer);
439 if (page_length != -1)
440 set_viewlength (page_length);
446 stc_custom_seed (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
448 lex_match (lexer, '=');
449 if (lex_match_id (lexer, "RANDOM"))
453 if (!lex_force_num (lexer))
455 set_rng (lex_number (lexer));
463 stc_custom_width (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
465 lex_match (lexer, '=');
466 if (lex_match_id (lexer, "NARROW"))
468 else if (lex_match_id (lexer, "WIDE"))
472 if (!lex_force_int (lexer))
474 if (lex_integer (lexer) < 40)
476 msg (SE, _("WIDTH must be at least 40."));
479 set_viewwidth (lex_integer (lexer));
486 /* Parses FORMAT subcommand, which consists of a numeric format
489 stc_custom_format (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
493 lex_match (lexer, '=');
494 if (!parse_format_specifier (lexer, &fmt))
496 if (fmt_is_string (fmt.type))
498 char str[FMT_STRING_LEN_MAX + 1];
499 msg (SE, _("FORMAT requires numeric output format as an argument. "
500 "Specified format %s is of type string."),
501 fmt_to_string (&fmt, str));
510 stc_custom_journal (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
512 lex_match (lexer, '=');
513 if (!lex_match_id (lexer, "ON") && !lex_match_id (lexer, "OFF"))
515 if (lex_token (lexer) == T_STRING)
519 lex_error (lexer, NULL);
527 stc_custom_listing (struct lexer *lexer, struct dataset *ds UNUSED, struct cmd_set *cmd UNUSED, void *aux UNUSED)
531 lex_match (lexer, '=');
532 if (lex_match_id (lexer, "ON") || lex_match_id (lexer, "YES"))
534 else if (lex_match_id (lexer, "OFF") || lex_match_id (lexer, "NO"))
541 outp_enable_device (listing, OUTP_DEV_LISTING);
547 stc_custom_disk (struct lexer *lexer, struct dataset *ds, struct cmd_set *cmd UNUSED, void *aux)
549 return stc_custom_listing (lexer, ds, cmd, aux);
553 show_blanks (const struct dataset *ds UNUSED)
555 if (get_blanks () == SYSMIS)
556 msg (SN, _("BLANKS is SYSMIS."));
558 msg (SN, _("BLANKS is %g."), get_blanks ());
563 format_cc (struct substring in, char grouping, char *out)
565 while (!ss_is_empty (in))
567 char c = ss_get_char (&in);
568 if (c == grouping || c == '\'')
578 show_cc (enum fmt_type type)
580 const struct fmt_number_style *cc = fmt_get_style (type);
581 char cc_string[FMT_STYLE_AFFIX_MAX * 4 * 2 + 3 + 1];
584 out = format_cc (cc->neg_prefix, cc->grouping, cc_string);
585 *out++ = cc->grouping;
586 out = format_cc (cc->prefix, cc->grouping, out);
587 *out++ = cc->grouping;
588 out = format_cc (cc->suffix, cc->grouping, out);
589 *out++ = cc->grouping;
590 out = format_cc (cc->neg_suffix, cc->grouping, out);
593 msg (SN, _("%s is \"%s\"."), fmt_name (type), cc_string);
597 show_cca (const struct dataset *ds UNUSED)
603 show_ccb (const struct dataset *ds UNUSED)
609 show_ccc (const struct dataset *ds UNUSED)
615 show_ccd (const struct dataset *ds UNUSED)
621 show_cce (const struct dataset *ds UNUSED)
627 show_decimals (const struct dataset *ds UNUSED)
629 msg (SN, _("DECIMAL is \"%c\"."), fmt_decimal_char (FMT_F));
633 show_endcmd (const struct dataset *ds UNUSED)
635 msg (SN, _("ENDCMD is \"%c\"."), get_endcmd ());
639 show_errors (const struct dataset *ds UNUSED)
641 bool terminal = get_error_routing_to_terminal ();
642 bool listing = get_error_routing_to_listing ();
643 msg (SN, _("ERRORS is \"%s\"."),
644 terminal && listing ? "BOTH"
645 : terminal ? "TERMINAL"
646 : listing ? "LISTING"
651 show_format (const struct dataset *ds UNUSED)
653 char str[FMT_STRING_LEN_MAX + 1];
654 msg (SN, _("FORMAT is %s."), fmt_to_string (get_format (), str));
658 show_length (const struct dataset *ds UNUSED)
660 msg (SN, _("LENGTH is %d."), get_viewlength ());
664 show_mxerrs (const struct dataset *ds UNUSED)
666 msg (SN, _("MXERRS is %d."), get_mxerrs ());
670 show_mxloops (const struct dataset *ds UNUSED)
672 msg (SN, _("MXLOOPS is %d."), get_mxloops ());
676 show_mxwarns (const struct dataset *ds UNUSED)
678 msg (SN, _("MXWARNS is %d."), get_mxwarns ());
681 /* Outputs that SETTING has the given INTEGER_FORMAT value. */
683 show_integer_format (const char *setting, enum integer_format integer_format)
685 msg (SN, _("%s is %s (%s)."),
687 (integer_format == INTEGER_MSB_FIRST ? "MSBFIRST"
688 : integer_format == INTEGER_LSB_FIRST ? "LSBFIRST"
690 integer_format == INTEGER_NATIVE ? "NATIVE" : "nonnative");
693 /* Outputs that SETTING has the given FLOAT_FORMAT value. */
695 show_float_format (const char *setting, enum float_format float_format)
697 const char *format_name = "";
699 switch (float_format)
701 case FLOAT_IEEE_SINGLE_LE:
702 format_name = "ISL (32-bit IEEE 754 single, little-endian)";
704 case FLOAT_IEEE_SINGLE_BE:
705 format_name = "ISB (32-bit IEEE 754 single, big-endian)";
707 case FLOAT_IEEE_DOUBLE_LE:
708 format_name = "IDL (64-bit IEEE 754 double, little-endian)";
710 case FLOAT_IEEE_DOUBLE_BE:
711 format_name = "IDB (64-bit IEEE 754 double, big-endian)";
715 format_name = "VF (32-bit VAX F, VAX-endian)";
718 format_name = "VD (64-bit VAX D, VAX-endian)";
721 format_name = "VG (64-bit VAX G, VAX-endian)";
725 format_name = "ZS (32-bit IBM Z hexadecimal short, big-endian)";
728 format_name = "ZL (64-bit IBM Z hexadecimal long, big-endian)";
736 msg (SN, _("%s is %s (%s)."),
737 setting, format_name,
738 float_format == FLOAT_NATIVE_DOUBLE ? "NATIVE" : "nonnative");
742 show_rib (const struct dataset *ds UNUSED)
744 show_integer_format ("RIB", data_in_get_integer_format ());
748 show_rrb (const struct dataset *ds UNUSED)
750 show_float_format ("RRB", data_in_get_float_format ());
754 show_scompression (const struct dataset *ds UNUSED)
756 if (get_scompression ())
757 msg (SN, _("SCOMPRESSION is ON."));
759 msg (SN, _("SCOMPRESSION is OFF."));
763 show_undefined (const struct dataset *ds UNUSED)
765 if (get_undefined ())
766 msg (SN, _("UNDEFINED is WARN."));
768 msg (SN, _("UNDEFINED is NOWARN."));
772 show_weight (const struct dataset *ds)
774 struct variable *var = dict_get_weight (dataset_dict (ds));
776 msg (SN, _("WEIGHT is off."));
778 msg (SN, _("WEIGHT is variable %s."), var_get_name (var));
782 show_wib (const struct dataset *ds UNUSED)
784 show_integer_format ("WIB", data_out_get_integer_format ());
788 show_wrb (const struct dataset *ds UNUSED)
790 show_float_format ("WRB", data_out_get_float_format ());
794 show_width (const struct dataset *ds UNUSED)
796 msg (SN, _("WIDTH is %d."), get_viewwidth ());
802 void (*function) (const struct dataset *);
805 const struct show_sbc show_table[] =
807 {"BLANKS", show_blanks},
813 {"DECIMALS", show_decimals},
814 {"ENDCMD", show_endcmd},
815 {"ERRORS", show_errors},
816 {"FORMAT", show_format},
817 {"LENGTH", show_length},
818 {"MXERRS", show_mxerrs},
819 {"MXLOOPS", show_mxloops},
820 {"MXWARNS", show_mxwarns},
823 {"SCOMPRESSION", show_scompression},
824 {"UNDEFINED", show_undefined},
825 {"WEIGHT", show_weight},
828 {"WIDTH", show_width},
832 show_all (const struct dataset *ds)
836 for (i = 0; i < sizeof show_table / sizeof *show_table; i++)
837 show_table[i].function (ds);
845 for (i = 0; i < 5; i++)
850 show_warranty (const struct dataset *ds UNUSED)
852 msg (MN, lack_of_warranty);
856 show_copying (const struct dataset *ds UNUSED)
862 cmd_show (struct lexer *lexer, struct dataset *ds)
864 if (lex_token (lexer) == '.')
872 if (lex_match (lexer, T_ALL))
874 else if (lex_match_id (lexer, "CC"))
876 else if (lex_match_id (lexer, "WARRANTY"))
878 else if (lex_match_id (lexer, "COPYING"))
880 else if (lex_token (lexer) == T_ID)
884 for (i = 0; i < sizeof show_table / sizeof *show_table; i++)
885 if (lex_match_id (lexer, show_table[i].name))
887 show_table[i].function (ds);
890 lex_error (lexer, NULL);
897 lex_error (lexer, NULL);
901 lex_match (lexer, '/');
903 while (lex_token (lexer) != '.');