1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2008 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU 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, see <http://www.gnu.org/licenses/>. */
19 #include <data/casereader-provider.h>
20 #include <libpspp/message.h>
21 #include <gl/xalloc.h>
22 #include <data/dictionary.h>
25 #include "psql-reader.h"
31 #include <libpspp/str.h>
34 #define _(msgid) gettext (msgid)
35 #define N_(msgid) (msgid)
40 psql_open_reader (struct psql_read_info *info, struct dictionary **dict)
42 msg (ME, _("Support for reading postgres databases was not compiled into this installation of PSPP"));
53 /* These macros must be the same as in catalog/pg_types.h from the postgres source */
66 #define BPCHAROID 1042
67 #define VARCHAROID 1043
70 #define TIMESTAMPOID 1114
71 #define TIMESTAMPTZOID 1184
72 #define INTERVALOID 1186
73 #define TIMETZOID 1266
74 #define NUMERICOID 1700
76 static void psql_casereader_destroy (struct casereader *reader UNUSED, void *r_);
78 static bool psql_casereader_read (struct casereader *, void *,
81 static const struct casereader_class psql_casereader_class =
84 psql_casereader_destroy,
95 bool integer_datetimes;
97 double postgres_epoch;
100 struct dictionary *dict;
102 /* An array of ints, which maps psql column numbers into
104 struct variable **vmap;
107 struct string fetch_cmd;
112 static bool set_value (struct psql_reader *r,
119 data_to_native (const void *in_, void *out_, int len)
122 const unsigned char *in = in_;
123 unsigned char *out = out_;
124 for (i = 0 ; i < len ; ++i )
129 data_to_native (const void *in_, void *out_, int len)
132 const unsigned char *in = in_;
133 unsigned char *out = out_;
134 for (i = 0 ; i < len ; ++i )
135 out[len - i - 1] = in[i];
140 #define GET_VALUE(IN, OUT) do { \
141 size_t sz = sizeof (OUT); \
142 data_to_native (*(IN), &(OUT), sz) ; \
149 dump (const unsigned char *x, int l)
153 for (i = 0; i < l ; ++i)
155 printf ("%02x ", x[i]);
160 for (i = 0; i < l ; ++i)
163 printf ("%c ", x[i]);
172 static struct variable *
173 create_var (struct psql_reader *r, const struct fmt_spec *fmt,
174 int width, const char *suggested_name, int col)
176 unsigned long int vx = 0;
177 struct variable *var;
178 char name[VAR_NAME_LEN + 1];
180 r->value_cnt += value_cnt_from_width (width);
182 if ( ! dict_make_unique_var_name (r->dict, suggested_name, &vx, name))
184 msg (ME, _("Cannot create variable name from %s"), suggested_name);
188 var = dict_create_var (r->dict, name, width);
189 var_set_both_formats (var, fmt);
193 r->vmap = xrealloc (r->vmap, (col + 1) * sizeof (*r->vmap));
196 r->vmapsize = col + 1;
207 reload_cache (struct psql_reader *r)
212 r->res = PQexec (r->conn, ds_cstr (&r->fetch_cmd));
214 if (PQresultStatus (r->res) != PGRES_TUPLES_OK || PQntuples (r->res) < 1)
226 psql_open_reader (struct psql_read_info *info, struct dictionary **dict)
229 int n_fields, n_tuples;
230 PGresult *qres = NULL;
231 casenumber n_cases = CASENUMBER_MAX;
233 struct psql_reader *r = xzalloc (sizeof *r);
234 struct string query ;
236 r->conn = PQconnectdb (info->conninfo);
237 if ( NULL == r->conn)
239 msg (ME, _("Memory error whilst opening psql source"));
243 if ( PQstatus (r->conn) != CONNECTION_OK )
245 msg (ME, _("Error opening psql source: %s."),
246 PQerrorMessage (r->conn));
253 const char *vers = PQparameterStatus (r->conn, "server_version");
255 sscanf (vers, "%d", &ver_num);
260 _("Postgres server is version %s."
261 " Reading from versions earlier than 8.0 is not supported."),
269 const char *dt = PQparameterStatus (r->conn, "integer_datetimes");
271 r->integer_datetimes = ( 0 == strcasecmp (dt, "on"));
275 if ( PQgetssl (r->conn) == NULL)
278 if (! info->allow_clear)
280 msg (ME, _("Connection is unencrypted, "
281 "but unencrypted connections have not been permitted."));
287 calendar_gregorian_to_offset (2000, 1, 1, NULL, NULL);
290 /* Create the dictionary and populate it */
291 *dict = r->dict = dict_create ();
294 select count (*) from (select * from medium) stupid_sql_standard;
297 ds_init_cstr (&query,
298 "BEGIN READ ONLY ISOLATION LEVEL SERIALIZABLE; "
299 "DECLARE pspp BINARY CURSOR FOR ");
301 ds_put_substring (&query, info->sql.ss);
303 qres = PQexec (r->conn, ds_cstr (&query));
305 if ( PQresultStatus (qres) != PGRES_COMMAND_OK )
307 msg (ME, _("Error from psql source: %s."),
308 PQresultErrorMessage (qres));
315 /* Now use the count() function to find the total number of cases
316 that this query returns.
317 Doing this incurs some overhead. The server has to iterate every
318 case in order to find this number. However, it's performed on the
319 server side, and in all except the most huge databases the extra
320 overhead will be worth the effort.
321 On the other hand, most PSPP functions don't need to know this.
322 The GUI is the notable exception.
324 ds_init_cstr (&query, "SELECT count (*) FROM (");
325 ds_put_substring (&query, info->sql.ss);
326 ds_put_cstr (&query, ") stupid_sql_standard");
328 qres = PQexec (r->conn, ds_cstr (&query));
330 if ( PQresultStatus (qres) != PGRES_TUPLES_OK )
332 msg (ME, _("Error from psql source: %s."),
333 PQresultErrorMessage (qres));
336 n_cases = atol (PQgetvalue (qres, 0, 0));
339 qres = PQexec (r->conn, "FETCH FIRST FROM pspp");
340 if ( PQresultStatus (qres) != PGRES_TUPLES_OK )
342 msg (ME, _("Error from psql source: %s."),
343 PQresultErrorMessage (qres));
347 n_tuples = PQntuples (qres);
348 n_fields = PQnfields (qres);
354 for (i = 0 ; i < n_fields ; ++i )
356 struct variable *var;
357 struct fmt_spec fmt = {FMT_F, 8, 2};
358 Oid type = PQftype (qres, i);
362 /* If there are no data then make a finger in the air
363 guess at the contents */
365 length = PQgetlength (qres, 0, i);
367 length = MAX_SHORT_STRING;
381 fmt.type = FMT_DOLLAR;
385 width = length > 0 ? length : 1;
393 width = (info->str_width == -1) ?
394 ROUND_UP (length, MAX_SHORT_STRING) : info->str_width;
400 width = length > 0 ? length : MAX_SHORT_STRING;
405 fmt.type = FMT_DTIME;
425 fmt.type = FMT_DATETIME;
437 msg (MW, _("Unsupported OID %d. SYSMIS values will be inserted."), type);
439 width = length > 0 ? length : MAX_SHORT_STRING;
445 var = create_var (r, &fmt, width, PQfname (qres, i), i);
446 if ( type == NUMERICOID && n_tuples > 0)
448 const uint8_t *vptr = (const uint8_t *) PQgetvalue (qres, 0, i);
450 int16_t n_digits, weight, dscale;
453 GET_VALUE (&vptr, n_digits);
454 GET_VALUE (&vptr, weight);
455 GET_VALUE (&vptr, sign);
456 GET_VALUE (&vptr, dscale);
460 fmt.w = fmt_max_output_width (fmt.type) ;
461 fmt.d = MIN (dscale, fmt_max_output_decimals (fmt.type, fmt.w));
462 var_set_both_formats (var, &fmt);
465 /* Timezones need an extra variable */
471 ds_init_cstr (&name, var_get_name (var));
472 ds_put_cstr (&name, "-zone");
477 create_var (r, &fmt, 0, ds_cstr (&name), -1);
486 ds_init_cstr (&name, var_get_name (var));
487 ds_put_cstr (&name, "-months");
492 create_var (r, &fmt, 0, ds_cstr (&name), -1);
503 qres = PQexec (r->conn, "MOVE BACKWARD 1 FROM pspp");
504 if ( PQresultStatus (qres) != PGRES_COMMAND_OK)
511 r->cache_size = info->bsize != -1 ? info->bsize: 4096;
513 ds_init_empty (&r->fetch_cmd);
514 ds_put_format (&r->fetch_cmd, "FETCH FORWARD %d FROM pspp", r->cache_size);
518 return casereader_create_sequential
522 &psql_casereader_class, r);
525 dict_destroy (*dict);
527 psql_casereader_destroy (NULL, r);
533 psql_casereader_destroy (struct casereader *reader UNUSED, void *r_)
535 struct psql_reader *r = r_;
539 ds_destroy (&r->fetch_cmd);
541 if (r->res) PQclear (r->res);
550 psql_casereader_read (struct casereader *reader UNUSED, void *r_,
553 struct psql_reader *r = r_;
555 if ( NULL == r->res || r->tuple >= r->cache_size)
557 if ( ! reload_cache (r) )
561 return set_value (r, cc);
565 set_value (struct psql_reader *r,
573 n_vars = PQnfields (r->res);
575 if ( r->tuple >= PQntuples (r->res))
578 case_create (c, r->value_cnt);
579 memset (case_data_rw_idx (c, 0)->s, ' ', MAX_SHORT_STRING * r->value_cnt);
582 for (i = 0 ; i < n_vars ; ++i )
584 Oid type = PQftype (r->res, i);
585 const struct variable *v = r->vmap[i];
586 union value *val = case_data_rw (c, v);
588 union value *val1 = NULL;
595 if (i < r->vmapsize && var_get_dict_index(v) + 1 < dict_get_var_cnt (r->dict))
597 const struct variable *v1 = NULL;
598 v1 = dict_get_var (r->dict, var_get_dict_index (v) + 1);
600 val1 = case_data_rw (c, v1);
608 if (PQgetisnull (r->res, r->tuple, i))
610 value_set_missing (val, var_get_width (v));
625 const uint8_t *vptr = (const uint8_t *) PQgetvalue (r->res, r->tuple, i);
626 int length = PQgetlength (r->res, r->tuple, i);
628 int var_width = var_get_width (v);
634 GET_VALUE (&vptr, x);
643 GET_VALUE (&vptr, x);
651 GET_VALUE (&vptr, x);
659 GET_VALUE (&vptr, x);
667 GET_VALUE (&vptr, n);
675 GET_VALUE (&vptr, n);
682 /* Postgres 8.3 uses 64 bits.
683 Earlier versions use 32 */
689 GET_VALUE (&vptr, x);
696 GET_VALUE (&vptr, x);
709 if ( r->integer_datetimes )
716 GET_VALUE (&vptr, things);
717 GET_VALUE (&vptr, us);
718 GET_VALUE (&vptr, days);
719 GET_VALUE (&vptr, months);
721 val->f = us / 1000000.0;
722 val->f += days * 24 * 3600;
728 uint32_t days, months;
731 GET_VALUE (&vptr, seconds);
732 GET_VALUE (&vptr, days);
733 GET_VALUE (&vptr, months);
736 val->f += days * 24 * 3600;
747 GET_VALUE (&vptr, x);
749 val->f = (x + r->postgres_epoch) * 24 * 3600 ;
755 if ( r->integer_datetimes)
758 GET_VALUE (&vptr, x);
759 val->f = x / 1000000.0;
764 GET_VALUE (&vptr, x);
773 if ( r->integer_datetimes)
778 GET_VALUE (&vptr, x);
779 val->f = x / 1000000.0;
785 GET_VALUE (&vptr, x);
789 GET_VALUE (&vptr, zone);
790 val1->f = zone / 3600.0;
797 if ( r->integer_datetimes)
801 GET_VALUE (&vptr, x);
805 val->f = (x + r->postgres_epoch * 24 * 3600 );
811 GET_VALUE (&vptr, x);
813 val->f = (x + r->postgres_epoch * 24 * 3600 );
821 memcpy (val->s, (char *) vptr, MIN (length, var_width));
828 int16_t n_digits, weight, dscale;
831 GET_VALUE (&vptr, n_digits);
832 GET_VALUE (&vptr, weight);
833 GET_VALUE (&vptr, sign);
834 GET_VALUE (&vptr, dscale);
841 fmt.w = fmt_max_output_width (fmt.type) ;
842 fmt.d = MIN (dscale, fmt_max_output_decimals (fmt.type, fmt.w));
843 var_set_both_formats (v, &fmt);
847 for (i = 0 ; i < n_digits; ++i)
850 GET_VALUE (&vptr, x);
851 f += x * pow (10000, weight--);