2 Frequently Asked Questions (FAQ) for PostgreSQL
4 Last updated: Sun Jan 12 13:36:11 EST 2003
6 Current maintainer: Bruce Momjian (pgman@candle.pha.pa.us)
8 The most recent version of this document can be viewed at
9 http://www.PostgreSQL.org/docs/faq-english.html.
11 Platform-specific questions are answered at
12 http://www.PostgreSQL.org/users-lounge/docs/faq.html.
13 _________________________________________________________________
17 1.1) What is PostgreSQL? How is it pronounced?
18 1.2) What is the copyright on PostgreSQL?
19 1.3) What Unix platforms does PostgreSQL run on?
20 1.4) What non-Unix ports are available?
21 1.5) Where can I get PostgreSQL?
22 1.6) Where can I get support?
23 1.7) What is the latest release?
24 1.8) What documentation is available?
25 1.9) How do I find out about known bugs or missing features?
26 1.10) How can I learn SQL?
27 1.11) Is PostgreSQL Y2K compliant?
28 1.12) How do I join the development team?
29 1.13) How do I submit a bug report?
30 1.14) How does PostgreSQL compare to other DBMSs?
31 1.15) How can I financially assist PostgreSQL?
35 2.1) Are there ODBC drivers for PostgreSQL?
36 2.2) What tools are available for using PostgreSQL with Web pages?
37 2.3) Does PostgreSQL have a graphical user interface?
38 2.4) What languages are available to communicate with PostgreSQL?
40 Administrative Questions
42 3.1) How do I install PostgreSQL somewhere other than
44 3.2) When I start postmaster, I get a Bad System Call or core dumped
46 3.3) When I try to start postmaster, I get IpcMemoryCreate errors.
48 3.4) When I try to start postmaster, I get IpcSemaphoreCreate errors.
50 3.5) How do I control connections from other hosts?
51 3.6) How do I tune the database engine for better performance?
52 3.7) What debugging features are available?
53 3.8) Why do I get "Sorry, too many clients" when trying to connect?
54 3.9) What is in the pgsql_tmp directory?
55 3.10) Why do I need to do a dump and restore to upgrade PostgreSQL
60 4.1) What is the difference between binary cursors and normal cursors?
61 4.2) How do I SELECT only the first few rows of a query?
62 4.3) How do I get a list of tables or other things I can see in psql?
63 4.4) How do you remove a column from a table?
64 4.5) What is the maximum size for a row, a table, and a database?
65 4.6) How much database disk space is required to store data from a
67 4.7) How do I find out what tables, indexes, databases, and users are
69 4.8) My queries are slow or don't make use of the indexes. Why?
70 4.9) How do I see how the query optimizer is evaluating my query?
71 4.10) What is an R-tree index?
72 4.11) What is the Genetic Query Optimizer?
73 4.12) How do I perform regular expression searches and
74 case-insensitive regular expression searches? How do I use an index
75 for case-insensitive searches?
76 4.13) In a query, how do I detect if a field is NULL?
77 4.14) What is the difference between the various character types?
78 4.15.1) How do I create a serial/auto-incrementing field?
79 4.15.2) How do I get the value of a SERIAL insert?
80 4.15.3) Don't currval() and nextval() lead to a race condition with
82 4.15.4) Why aren't my sequence numbers reused on transaction abort?
83 Why are there gaps in the numbering of my sequence/SERIAL column?
84 4.16) What is an OID? What is a TID?
85 4.17) What is the meaning of some of the terms used in PostgreSQL?
86 4.18) Why do I get the error "ERROR: Memory exhausted in
88 4.19) How do I tell what PostgreSQL version I am running?
89 4.20) Why does my large-object operations get "invalid large obj
91 4.21) How do I create a column that will default to the current time?
92 4.22) Why are my subqueries using IN so slow?
93 4.23) How do I perform an outer join?
94 4.24) How do I perform queries using multiple databases?
95 4.25) How do I return multiple rows or columns from a function?
96 4.26) Why can't I reliably create/drop temporary tables in PL/PgSQL
98 4.27) What replication options are available?
99 4.28) What encryption options are available?
103 5.1) I wrote a user-defined function. When I run it in psql, why does
105 5.2) How can I contribute some nifty new types and functions to
107 5.3) How do I write a C function to return a tuple?
108 5.4) I have changed a source file. Why does the recompile not see the
110 _________________________________________________________________
114 1.1) What is PostgreSQL? How is it pronounced?
116 PostgreSQL is pronounced Post-Gres-Q-L.
118 PostgreSQL is an enhancement of the POSTGRES database management
119 system, a next-generation DBMS research prototype. While PostgreSQL
120 retains the powerful data model and rich data types of POSTGRES, it
121 replaces the PostQuel query language with an extended subset of SQL.
122 PostgreSQL is free and the complete source is available.
124 PostgreSQL development is performed by a team of developers who all
125 subscribe to the PostgreSQL development mailing list. The current
126 coordinator is Marc G. Fournier (scrappy@PostgreSQL.org). (See section
127 1.6 on how to join). This team is now responsible for all development
130 The authors of PostgreSQL 1.01 were Andrew Yu and Jolly Chen. Many
131 others have contributed to the porting, testing, debugging, and
132 enhancement of the code. The original Postgres code, from which
133 PostgreSQL is derived, was the effort of many graduate students,
134 undergraduate students, and staff programmers working under the
135 direction of Professor Michael Stonebraker at the University of
136 California, Berkeley.
138 The original name of the software at Berkeley was Postgres. When SQL
139 functionality was added in 1995, its name was changed to Postgres95.
140 The name was changed at the end of 1996 to PostgreSQL.
142 1.2) What is the copyright on PostgreSQL?
144 PostgreSQL is subject to the following COPYRIGHT:
146 PostgreSQL Data Base Management System
148 Portions copyright (c) 1996-2002, PostgreSQL Global Development Group
149 Portions Copyright (c) 1994-6 Regents of the University of California
151 Permission to use, copy, modify, and distribute this software and its
152 documentation for any purpose, without fee, and without a written
153 agreement is hereby granted, provided that the above copyright notice
154 and this paragraph and the following two paragraphs appear in all
157 IN NO EVENT SHALL THE UNIVERSITY OF CALIFORNIA BE LIABLE TO ANY PARTY
158 FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES,
159 INCLUDING LOST PROFITS, ARISING OUT OF THE USE OF THIS SOFTWARE AND
160 ITS DOCUMENTATION, EVEN IF THE UNIVERSITY OF CALIFORNIA HAS BEEN
161 ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
163 THE UNIVERSITY OF CALIFORNIA SPECIFICALLY DISCLAIMS ANY WARRANTIES,
164 INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
165 MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE
166 PROVIDED HEREUNDER IS ON AN "AS IS" BASIS, AND THE UNIVERSITY OF
167 CALIFORNIA HAS NO OBLIGATIONS TO PROVIDE MAINTENANCE, SUPPORT,
168 UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
170 The above is the BSD license, the classic open-source license. It has
171 no restrictions on how the source code may be used. We like it and
172 have no intention of changing it.
174 1.3) What Unix platforms does PostgreSQL run on?
176 In general, a modern Unix-compatible platform should be able to run
177 PostgreSQL. The platforms that had received explicit testing at the
178 time of release are listed in the installation instructions.
180 1.4) What non-Unix ports are available?
184 It is possible to compile the libpq C library, psql, and other
185 interfaces and client applications to run on MS Windows platforms. In
186 this case, the client is running on MS Windows, and communicates via
187 TCP/IP to a server running on one of our supported Unix platforms. A
188 file win32.mak is included in the distribution for making a Win32
189 libpq library and psql. PostgreSQL also communicates with ODBC
194 The database server can run on Windows NT and Win2k using Cygwin, the
195 Cygnus Unix/NT porting library. See pgsql/doc/FAQ_MSWIN in the
196 distribution or the MS Windows FAQ at
197 http://www.PostgreSQL.org/docs/faq-mswin.html.
199 A native port to MS Win NT/2000/XP is currently being worked on.
201 1.5) Where can I get PostgreSQL?
203 The primary anonymous ftp site for PostgreSQL is
204 ftp://ftp.PostgreSQL.org/pub. For mirror sites, see our main web site.
206 1.6) Where can I get support?
208 The main mailing list is: pgsql-general@PostgreSQL.org. It is
209 available for discussion of matters pertaining to PostgreSQL. To
210 subscribe, send mail with the following lines in the body (not the
215 to pgsql-general-request@PostgreSQL.org.
217 There is also a digest list available. To subscribe to this list, send
218 email to: pgsql-general-digest-request@PostgreSQL.org with a body of:
222 Digests are sent out to members of this list whenever the main list
223 has received around 30k of messages.
225 The bugs mailing list is available. To subscribe to this list, send
226 email to pgsql-bugs-request@PostgreSQL.org with a body of:
230 There is also a developers discussion mailing list available. To
231 subscribe to this list, send email to
232 pgsql-hackers-request@PostgreSQL.org with a body of:
236 Additional mailing lists and information about PostgreSQL can be found
237 via the PostgreSQL WWW home page at:
239 http://www.PostgreSQL.org
241 There is also an IRC channel on EFNet, channel #PostgreSQL. I use the
242 Unix command irc -c '#PostgreSQL' "$USER" irc.phoenix.net.
244 A list of commercial support companies is available at
245 http://www.PostgreSQL.org/users-lounge/commercial-support.html.
247 1.7) What is the latest release?
249 The latest release of PostgreSQL is version 7.3.1.
251 We plan to have major releases every four months.
253 1.8) What documentation is available?
255 Several manuals, manual pages, and some small test examples are
256 included in the distribution. See the /doc directory. You can also
257 browse the manual online at
258 http://www.PostgreSQL.org/users-lounge/docs/.
260 There are two PostgreSQL books available online at
261 http://www.PostgreSQL.org/docs/awbook.html and
262 http://www.commandprompt.com/ppbook/. There is a list of PostgreSQL
263 books available for purchase at http://www.ca.PostgreSQL.org/books/.
264 There is also a collection of PostgreSQL technical articles at
265 http://techdocs.PostgreSQL.org/.
267 psql has some nice \d commands to show information about types,
268 operators, functions, aggregates, etc.
270 Our web site contains even more documentation.
272 1.9) How do I find out about known bugs or missing features?
274 PostgreSQL supports an extended subset of SQL-92. See our TODO list
275 for known bugs, missing features, and future plans.
277 1.10) How can I learn SQL?
279 The PostgreSQL book at http://www.PostgreSQL.org/docs/awbook.html
280 teaches SQL. There is another PostgreSQL book at
281 http://www.commandprompt.com/ppbook. There is a nice tutorial at
282 http://www.intermedia.net/support/sql/sqltut.shtm, at
283 http://ourworld.compuserve.com/homepages/graeme_birchall/HTM_COOK.HTM,
284 and at http://sqlcourse.com.
286 Another one is "Teach Yourself SQL in 21 Days, Second Edition" at
287 http://members.tripod.com/er4ebus/sql/index.htm
289 Many of our users like The Practical SQL Handbook, Bowman, Judith S.,
290 et al., Addison-Wesley. Others like The Complete Reference SQL, Groff
293 1.11) Is PostgreSQL Y2K compliant?
295 Yes, we easily handle dates past the year 2000 AD, and before 2000 BC.
297 1.12) How do I join the development team?
299 First, download the latest source and read the PostgreSQL Developers
300 documentation on our web site, or in the distribution. Second,
301 subscribe to the pgsql-hackers and pgsql-patches mailing lists. Third,
302 submit high quality patches to pgsql-patches.
304 There are about a dozen people who have commit privileges to the
305 PostgreSQL CVS archive. They each have submitted so many high-quality
306 patches that it was impossible for the existing committers to keep up,
307 and we had confidence that patches they committed were of high
310 1.13) How do I submit a bug report?
312 Please visit the PostgreSQL BugTool page at
313 http://www.PostgreSQL.org/bugs/bugs.php, which gives guidelines and
314 directions on how to submit a bug report.
316 Also check out our ftp site ftp://ftp.PostgreSQL.org/pub to see if
317 there is a more recent PostgreSQL version or patches.
319 1.14) How does PostgreSQL compare to other DBMSs?
321 There are several ways of measuring software: features, performance,
322 reliability, support, and price.
325 PostgreSQL has most features present in large commercial DBMSs,
326 like transactions, subselects, triggers, views, foreign key
327 referential integrity, and sophisticated locking. We have some
328 features they do not have, like user-defined types,
329 inheritance, rules, and multi-version concurrency control to
330 reduce lock contention.
333 PostgreSQL has performance similar to other commercial and open
334 source databases. it is faster for some things, slower for
335 others. In comparison to MySQL or leaner database systems, we
336 are slower on inserts/updates because of transaction overhead.
337 Of course, MySQL does not have any of the features mentioned in
338 the Features section above. We are built for reliability and
339 features, though we continue to improve performance in every
340 release. There is an interesting Web page comparing PostgreSQL
341 to MySQL at http://openacs.org/philosophy/why-not-mysql.html
344 We realize that a DBMS must be reliable, or it is worthless. We
345 strive to release well-tested, stable code that has a minimum
346 of bugs. Each release has at least one month of beta testing,
347 and our release history shows that we can provide stable, solid
348 releases that are ready for production use. We believe we
349 compare favorably to other database software in this area.
352 Our mailing lists provide contact with a large group of
353 developers and users to help resolve any problems encountered.
354 While we cannot guarantee a fix, commercial DBMSs do not always
355 supply a fix either. Direct access to developers, the user
356 community, manuals, and the source code often make PostgreSQL
357 support superior to other DBMSs. There is commercial
358 per-incident support available for those who need it. (See FAQ
362 We are free for all use, both commercial and non-commercial.
363 You can add our code to your product with no limitations,
364 except those outlined in our BSD-style license stated above.
366 1.15) How can I financially assist PostgreSQL?
368 PostgreSQL has had a first-class infrastructure since we started in
369 1996. This is all thanks to Marc Fournier, who has created and managed
370 this infrastructure over the years.
372 Quality infrastructure is very important to an open-source project. It
373 prevents disruptions that can greatly delay forward movement of the
376 Of course, this infrastructure is not cheap. There are a variety of
377 monthly and one-time expenses that are required to keep it going. If
378 you or your company has money it can donate to help fund this effort,
379 please go to http://store.pgsql.com/shopping/ and make a donation.
381 Although the web page mentions PostgreSQL, Inc, the "contributions"
382 item is solely to support the PostgreSQL project and does not fund any
383 specific company. If you prefer, you can also send a check to the
385 _________________________________________________________________
387 User Client Questions
389 2.1) Are there ODBC drivers for PostgreSQL?
391 There are two ODBC drivers available, PsqlODBC and OpenLink ODBC.
393 You can download PsqlODBC from
394 http://gborg.postgresql.org/project/psqlodbc/projdisplay.php.
396 OpenLink ODBC can be gotten from http://www.openlinksw.com. It works
397 with their standard ODBC client software so you'll have PostgreSQL
398 ODBC available on every client platform they support (Win, Mac, Unix,
401 They will probably be selling this product to people who need
402 commercial-quality support, but a freeware version will always be
403 available. Please send questions to postgres95@openlink.co.uk.
405 2.2) What tools are available for using PostgreSQL with Web pages?
407 A nice introduction to Database-backed Web pages can be seen at:
408 http://www.webreview.com
410 For Web integration, PHP is an excellent interface. It is at
413 For complex cases, many use the Perl interface and CGI.pm or mod_perl.
415 2.3) Does PostgreSQL have a graphical user interface?
417 Yes, there are several graphical interfaces to PostgreSQL available.
418 These include PgAccess http://www.pgaccess.org), PgAdmin II
419 (http://www.pgadmin.org, Win32-only), RHDB Admin
420 (http://sources.redhat.com/rhdb/ ) and Rekall (
421 http://www.thekompany.com/products/rekall/, proprietary). There is
422 also PHPPgAdmin ( http://phppgadmin.sourceforge.net/ ), a web-based
423 interface to PostgreSQL.
425 2.4) What languages are able to communicate with PostgreSQL?
427 Most popular programming languages contain an interface to PostgreSQL.
428 Check your programming language's list of extension modules.
430 The following interfaces are included in the PostgreSQL distribution:
437 Additional interfaces are available at http://gborg.postgresql.org in
438 the Drivers/Interfaces section.
439 _________________________________________________________________
441 Administrative Questions
443 3.1) How do I install PostgreSQL somewhere other than /usr/local/pgsql?
445 Specify the --prefix option when running configure.
447 3.2) When I start postmaster, I get a Bad System Call or core dumped
450 It could be a variety of problems, but first check to see that you
451 have System V extensions installed in your kernel. PostgreSQL requires
452 kernel support for shared memory and semaphores.
454 3.3) When I try to start postmaster, I get IpcMemoryCreate errors. Why?
456 You either do not have shared memory configured properly in your
457 kernel or you need to enlarge the shared memory available in the
458 kernel. The exact amount you need depends on your architecture and how
459 many buffers and backend processes you configure for postmaster. For
460 most systems, with default numbers of buffers and processes, you need
461 a minimum of ~1 MB. See the PostgreSQL Administrator's Guide for more
462 detailed information about shared memory and semaphores.
464 3.4) When I try to start postmaster, I get IpcSemaphoreCreate errors. Why?
466 If the error message is IpcSemaphoreCreate: semget failed (No space
467 left on device) then your kernel is not configured with enough
468 semaphores. Postgres needs one semaphore per potential backend
469 process. A temporary solution is to start postmaster with a smaller
470 limit on the number of backend processes. Use -N with a parameter less
471 than the default of 32. A more permanent solution is to increase your
472 kernel's SEMMNS and SEMMNI parameters.
474 Inoperative semaphores can also cause crashes during heavy database
477 If the error message is something else, you might not have semaphore
478 support configured in your kernel at all. See the PostgreSQL
479 Administrator's Guide for more detailed information about shared
480 memory and semaphores.
482 3.5) How do I control connections from other hosts?
484 By default, PostgreSQL only allows connections from the local machine
485 using Unix domain sockets. Other machines will not be able to connect
486 unless you add the -i flag to postmaster, and enable host-based
487 authentication by modifying the file $PGDATA/pg_hba.conf accordingly.
488 This will allow TCP/IP connections.
490 3.6) How do I tune the database engine for better performance?
492 Certainly, indexes can speed up queries. The EXPLAIN command allows
493 you to see how PostgreSQL is interpreting your query, and which
494 indexes are being used.
496 If you are doing many INSERTs, consider doing them in a large batch
497 using the COPY command. This is much faster than individual INSERTS.
498 Second, statements not in a BEGIN WORK/COMMIT transaction block are
499 considered to be in their own transaction. Consider performing several
500 statements in a single transaction block. This reduces the transaction
501 overhead. Also, consider dropping and recreating indexes when making
504 There are several tuning options. You can disable fsync() by starting
505 postmaster with a -o -F option. This will prevent fsync()s from
506 flushing to disk after every transaction.
508 You can also use the postmaster -B option to increase the number of
509 shared memory buffers used by the backend processes. If you make this
510 parameter too high, the postmaster may not start because you have
511 exceeded your kernel's limit on shared memory space. Each buffer is 8K
512 and the default is 64 buffers.
514 You can also use the backend -S option to increase the maximum amount
515 of memory used by the backend process for temporary sorts. The -S
516 value is measured in kilobytes, and the default is 512 (i.e. 512K).
518 You can also use the CLUSTER command to group data in tables to match
519 an index. See the CLUSTER manual page for more details.
521 3.7) What debugging features are available?
523 PostgreSQL has several features that report status information that
524 can be valuable for debugging purposes.
526 First, by running configure with the --enable-cassert option, many
527 assert()s monitor the progress of the backend and halt the program
528 when something unexpected occurs.
530 Both postmaster and postgres have several debug options available.
531 First, whenever you start postmaster, make sure you send the standard
532 output and error to a log file, like:
534 ./bin/postmaster >server.log 2>&1 &
536 This will put a server.log file in the top-level PostgreSQL directory.
537 This file contains useful information about problems or errors
538 encountered by the server. Postmaster has a -d option that allows even
539 more detailed information to be reported. The -d option takes a number
540 that specifies the debug level. Be warned that high debug level values
541 generate large log files.
543 If postmaster is not running, you can actually run the postgres
544 backend from the command line, and type your SQL statement directly.
545 This is recommended only for debugging purposes. Note that a newline
546 terminates the query, not a semicolon. If you have compiled with
547 debugging symbols, you can use a debugger to see what is happening.
548 Because the backend was not started from postmaster, it is not running
549 in an identical environment and locking/backend interaction problems
550 may not be duplicated.
552 If postmaster is running, start psql in one window, then find the PID
553 of the postgres process used by psql. Use a debugger to attach to the
554 postgres PID. You can set breakpoints in the debugger and issue
555 queries from psql. If you are debugging postgres startup, you can set
556 PGOPTIONS="-W n", then start psql. This will cause startup to delay
557 for n seconds so you can attach to the process with the debugger, set
558 any breakpoints, and continue through the startup sequence.
560 The postgres program has -s, -A, and -t options that can be very
561 useful for debugging and performance measurements.
563 You can also compile with profiling to see what functions are taking
564 execution time. The backend profile files will be deposited in the
565 pgsql/data/base/dbname directory. The client profile file will be put
566 in the client's current directory. Linux requires a compile with
567 -DLINUX_PROFILE for proper profiling.
569 3.8) Why do I get "Sorry, too many clients" when trying to connect?
571 You need to increase postmaster's limit on how many concurrent backend
572 processes it can start.
574 The default limit is 32 processes. You can increase it by restarting
575 postmaster with a suitable -N value or modifying postgresql.conf.
577 Note that if you make -N larger than 32, you must also increase -B
578 beyond its default of 64; -B must be at least twice -N, and probably
579 should be more than that for best performance. For large numbers of
580 backend processes, you are also likely to find that you need to
581 increase various Unix kernel configuration parameters. Things to check
582 include the maximum size of shared memory blocks, SHMMAX; the maximum
583 number of semaphores, SEMMNS and SEMMNI; the maximum number of
584 processes, NPROC; the maximum number of processes per user, MAXUPRC;
585 and the maximum number of open files, NFILE and NINODE. The reason
586 that PostgreSQL has a limit on the number of allowed backend processes
587 is so your system won't run out of resources.
589 In PostgreSQL versions prior to 6.5, the maximum number of backends
590 was 64, and changing it required a rebuild after altering the
591 MaxBackendId constant in include/storage/sinvaladt.h.
593 3.9) What is in the pgsql_tmp directory?
595 This directory contains temporary files generated by the query
596 executor. For example, if a sort needs to be done to satisfy an ORDER
597 BY and the sort requires more space than the backend's -S parameter
598 allows, then temporary files are created here to hold the extra data.
600 The temporary files are usually deleted automatically, but might
601 remain if a backend crashes during a sort. A stop and restart of the
602 postmaster will remove files from those directories.
604 3.10) Why do I need to do a dump and restore to upgrade between major
607 The PostgreSQL team makes only small changes between minor releases,
608 so upgrading from 7.2 to 7.2.1 does not require a dump and restore.
609 However, major releases (e.g. from 7.2 to 7.3) often change the
610 internal format of system tables and data files. These changes are
611 often complex, so we don't maintain backward compatability for data
612 files. A dump outputs data in a generic format that can then be loaded
613 in using the new internal format.
615 In releases where the on-disk format does not change, the pg_upgrade
616 script can be used to upgrade without a dump/restore. The release
617 notes mention whether pg_upgrade is available for the release.
618 _________________________________________________________________
620 Operational Questions
622 4.1) What is the difference between binary cursors and normal cursors?
624 See the DECLARE manual page for a description.
626 4.2) How do I SELECT only the first few rows of a query?
628 See the FETCH manual page, or use SELECT ... LIMIT....
630 The entire query may have to be evaluated, even if you only want the
631 first few rows. Consider using a query that has an ORDER BY. If there
632 is an index that matches the ORDER BY, PostgreSQL may be able to
633 evaluate only the first few records requested, or the entire query may
634 have to be evaluated until the desired rows have been generated.
636 4.3) How do I get a list of tables or other things I can see in psql?
638 You can read the source code for psql in file
639 pgsql/src/bin/psql/describe.c. It contains SQL commands that generate
640 the output for psql's backslash commands. You can also start psql with
641 the -E option so it will print out the queries it uses to execute the
644 4.4) How do you remove a column from a table?
646 This functionality was added in release 7.3 with ALTER TABLE DROP
647 COLUMN. In earlier versions, you can do this:
649 LOCK TABLE old_table;
650 SELECT ... -- select all columns but the one you want to remove
653 DROP TABLE old_table;
654 ALTER TABLE new_table RENAME TO old_table;
657 4.5) What is the maximum size for a row, a table, and a database?
659 These are the limits:
660 Maximum size for a database? unlimited (1 TB databases exist)
661 Maximum size for a table? 16 TB
662 Maximum size for a row? 1.6TB
663 Maximum size for a field? 1 GB
664 Maximum number of rows in a table? unlimited
665 Maximum number of columns in a table? 250-1600 depending on column types
666 Maximum number of indexes on a table? unlimited
668 Of course, these are not actually unlimited, but limited to available
669 disk space and memory/swap space. Performance may suffer when these
670 values get unusually large.
672 The maximum table size of 16 TB does not require large file support
673 from the operating system. Large tables are stored as multiple 1 GB
674 files so file system size limits are not important.
676 The maximum table size and maximum number of columns can be increased
677 if the default block size is increased to 32k.
679 4.6) How much database disk space is required to store data from a typical
682 A PostgreSQL database may require up to five times the disk space to
683 store data from a text file.
685 As an example, consider a file of 100,000 lines with an integer and
686 text description on each line. Suppose the text string avergages
687 twenty bytes in length. The flat file would be 2.8 MB. The size of the
688 PostgreSQL database file containing this data can be estimated as 6.4
690 36 bytes: each row header (approximate)
691 24 bytes: one int field and one text field
692 + 4 bytes: pointer on page to tuple
693 ----------------------------------------
696 The data page size in PostgreSQL is 8192 bytes (8 KB), so:
699 ------------------- = 128 rows per database page (rounded down)
703 -------------------- = 782 database pages (rounded up)
706 782 database pages * 8192 bytes per page = 6,406,144 bytes (6.4 MB)
708 Indexes do not require as much overhead, but do contain the data that
709 is being indexed, so they can be large also.
711 NULLs are stored in bitmaps, so they use very little space.
713 4.7) How do I find out what tables, indexes, databases, and users are
716 psql has a variety of backslash commands to show such information. Use
717 \? to see them. There are also system tables beginning with pg_ that
718 describe these too. Also, psql -l will list all databases.
720 Also try the file pgsql/src/tutorial/syscat.source. It illustrates
721 many of the SELECTs needed to get information from the database system
724 4.8) My queries are slow or don't make use of the indexes. Why?
726 Indexes are not automatically used by every query. Indexes are only
727 used if the table is larger than a minimum size, and the query selects
728 only a small percentage of the rows in the table. This is because the
729 random disk access caused by an index scan can be slower than a
730 straight read through the table, or sequential scan.
732 To determine if an index should be used, PostgreSQL must have
733 statistics about the table. These statistics are collected using
734 VACUUM ANALYZE, or simply ANALYZE. Using statistics, the optimizer
735 knows how many rows are in the table, and can better determine if
736 indexes should be used. Statistics are also valuable in determining
737 optimal join order and join methods. Statistics collection should be
738 performed periodically as the contents of the table change.
740 Indexes are normally not used for ORDER BY or to perform joins. A
741 sequential scan followed by an explicit sort is usually faster than an
742 index scan of a large table.
743 However, LIMIT combined with ORDER BY often will use an index because
744 only a small portion of the table is returned. In fact, though MAX()
745 and MIN() don't use indexes, it is possible to retrieve such values
746 using an index with ORDER BY and LIMIT:
749 ORDER BY col [ DESC ]
752 When using wild-card operators such as LIKE or ~, indexes can only be
753 used in certain circumstances:
754 * The beginning of the search string must be anchored to the start
756 + LIKE patterns must not start with %.
757 + ~ (regular expression) patterns must start with ^.
758 * The search string can not start with a character class, e.g.
760 * Case-insensitive searches such as ILIKE and ~* do not utilise
761 indexes. Instead, use functional indexes, which are described in
763 * The default C locale must be used during initdb.
765 4.9) How do I see how the query optimizer is evaluating my query?
767 See the EXPLAIN manual page.
769 4.10) What is an R-tree index?
771 An R-tree index is used for indexing spatial data. A hash index can't
772 handle range searches. A B-tree index only handles range searches in a
773 single dimension. R-trees can handle multi-dimensional data. For
774 example, if an R-tree index can be built on an attribute of type
775 point, the system can more efficiently answer queries such as "select
776 all points within a bounding rectangle."
778 The canonical paper that describes the original R-tree design is:
780 Guttman, A. "R-trees: A Dynamic Index Structure for Spatial
781 Searching." Proceedings of the 1984 ACM SIGMOD Int'l Conf on Mgmt of
784 You can also find this paper in Stonebraker's "Readings in Database
787 Built-in R-trees can handle polygons and boxes. In theory, R-trees can
788 be extended to handle higher number of dimensions. In practice,
789 extending R-trees requires a bit of work and we don't currently have
790 any documentation on how to do it.
792 4.11) What is the Genetic Query Optimizer?
794 The GEQO module speeds query optimization when joining many tables by
795 means of a Genetic Algorithm (GA). It allows the handling of large
796 join queries through nonexhaustive search.
798 4.12) How do I perform regular expression searches and case-insensitive
799 regular expression searches? How do I use an index for case-insensitive
802 The ~ operator does regular expression matching, and ~* does
803 case-insensitive regular expression matching. The case-insensitive
804 variant of LIKE is called ILIKE.
806 Case-insensitive equality comparisons are normally expressed as:
809 WHERE lower(col) = 'abc';
811 This will not use an standard index. However, if you create a
812 functional index, it will be used:
813 CREATE INDEX tabindex ON tab (lower(col));
815 4.13) In a query, how do I detect if a field is NULL?
817 You test the column with IS NULL and IS NOT NULL.
819 4.14) What is the difference between the various character types?
821 Type Internal Name Notes
822 --------------------------------------------------
823 VARCHAR(n) varchar size specifies maximum length, no padding
824 CHAR(n) bpchar blank padded to the specified fixed length
825 TEXT text no specific upper limit on length
826 "char" char one character
827 BYTEA bytea variable-length byte array (null-byte safe)
829 You will see the internal name when examining system catalogs and in
832 The last four types above are "varlena" types (i.e., the first four
833 bytes on disk are the length, followed by the data). Thus the actual
834 space used is slightly greater than the declared size. However, these
835 data types are also subject to compression or being stored out-of-line
836 by TOAST, so the space on disk might also be less than expected.
837 VARCHAR(n) is best when storing variable-length strings and it limits
838 how long a string can be. TEXT is for strings of unlimited length,
839 with a maximum of one gigabyte.
841 CHAR(n) is for storing strings that are all the same length. CHAR(n)
842 pads with blanks to the specified length, while VARCHAR(n) only stores
843 the characters supplied. BYTEA is for storing binary data,
844 particularly values that include NULL bytes. These types have similar
845 performance characteristics.
847 4.15.1) How do I create a serial/auto-incrementing field?
849 PostgreSQL supports a SERIAL data type. It auto-creates a sequence and
850 index on the column. For example, this:
851 CREATE TABLE person (
856 is automatically translated into this:
857 CREATE SEQUENCE person_id_seq;
858 CREATE TABLE person (
859 id INT4 NOT NULL DEFAULT nextval('person_id_seq'),
862 CREATE UNIQUE INDEX person_id_key ON person ( id );
864 See the create_sequence manual page for more information about
865 sequences. You can also use each row's OID field as a unique value.
866 However, if you need to dump and reload the database, you need to use
867 pg_dump's -o option or COPY WITH OIDS option to preserve the OIDs.
869 4.15.2) How do I get the value of a SERIAL insert?
871 One approach is to retrieve the next SERIAL value from the sequence
872 object with the nextval() function before inserting and then insert it
873 explicitly. Using the example table in 4.15.1, an example in a
874 pseudo-language would look like this:
875 new_id = execute("SELECT nextval('person_id_seq')");
876 execute("INSERT INTO person (id, name) VALUES (new_id, 'Blaise Pascal')");
878 You would then also have the new value stored in new_id for use in
879 other queries (e.g., as a foreign key to the person table). Note that
880 the name of the automatically created SEQUENCE object will be named
881 <table>_<serialcolumn>_seq, where table and serialcolumn are the names
882 of your table and your SERIAL column, respectively.
884 Alternatively, you could retrieve the assigned SERIAL value with the
885 currval() function after it was inserted by default, e.g.,
886 execute("INSERT INTO person (name) VALUES ('Blaise Pascal')");
887 new_id = execute("SELECT currval('person_id_seq')");
889 Finally, you could use the OID returned from the INSERT statement to
890 look up the default value, though this is probably the least portable
891 approach. In Perl, using DBI with Edmund Mergl's DBD::Pg module, the
892 oid value is made available via $sth->{pg_oid_status} after
895 4.15.3) Don't currval() and nextval() lead to a race condition with other
898 No. currval() returns the current value assigned by your backend, not
901 4.15.4) Why aren't my sequence numbers reused on transaction abort? Why are
902 there gaps in the numbering of my sequence/SERIAL column?
904 To improve concurrency, sequence values are given out to running
905 transactions as needed and are not locked until the transaction
906 completes. This causes gaps in numbering from aborted transactions.
908 4.16) What is an OID? What is a TID?
910 OIDs are PostgreSQL's answer to unique row ids. Every row that is
911 created in PostgreSQL gets a unique OID. All OIDs generated during
912 initdb are less than 16384 (from include/access/transam.h). All
913 user-created OIDs are equal to or greater than this. By default, all
914 these OIDs are unique not only within a table or database, but unique
915 within the entire PostgreSQL installation.
917 PostgreSQL uses OIDs in its internal system tables to link rows
918 between tables. These OIDs can be used to identify specific user rows
919 and used in joins. It is recommended you use column type OID to store
920 OID values. You can create an index on the OID field for faster
923 OIDs are assigned to all new rows from a central area that is used by
924 all databases. If you want to change the OID to something else, or if
925 you want to make a copy of the table, with the original OIDs, there is
926 no reason you can't do it:
927 CREATE TABLE new_table(old_oid oid, mycol int);
928 SELECT old_oid, mycol INTO new FROM old;
929 COPY new TO '/tmp/pgtable';
931 COPY new WITH OIDS FROM '/tmp/pgtable';
933 OIDs are stored as 4-byte integers, and will overflow at 4 billion. No
934 one has reported this ever happening, and we plan to have the limit
935 removed before anyone does.
937 TIDs are used to identify specific physical rows with block and offset
938 values. TIDs change after rows are modified or reloaded. They are used
939 by index entries to point to physical rows.
941 4.17) What is the meaning of some of the terms used in PostgreSQL?
943 Some of the source code and older documentation use terms that have
944 more common usage. Here are some:
945 * table, relation, class
947 * column, field, attribute
953 * range variable, table name, table alias
955 A list of general database terms can be found at:
956 http://hea-www.harvard.edu/MST/simul/software/docs/pkgs/pgsql/glossary
959 4.18) Why do I get the error "ERROR: Memory exhausted in AllocSetAlloc()"?
961 You probably have run out of virtual memory on your system, or your
962 kernel has a low limit for certain resources. Try this before starting
967 Depending on your shell, only one of these may succeed, but it will
968 set your process data segment limit much higher and perhaps allow the
969 query to complete. This command applies to the current process, and
970 all subprocesses created after the command is run. If you are having a
971 problem with the SQL client because the backend is returning too much
972 data, try it before starting the client.
974 4.19) How do I tell what PostgreSQL version I am running?
976 From psql, type SELECT version();
978 4.20) Why does my large-object operations get "invalid large obj
981 You need to put BEGIN WORK and COMMIT around any use of a large object
982 handle, that is, surrounding lo_open ... lo_close.
984 Currently PostgreSQL enforces the rule by closing large object handles
985 at transaction commit. So the first attempt to do anything with the
986 handle will draw invalid large obj descriptor. So code that used to
987 work (at least most of the time) will now generate that error message
988 if you fail to use a transaction.
990 If you are using a client interface like ODBC you may need to set
993 4.21) How do I create a column that will default to the current time?
995 Use CURRENT_TIMESTAMP:
996 CREATE TABLE test (x int, modtime timestamp DEFAULT CURRENT_TIMESTAMP );
998 4.22) Why are my subqueries using IN so slow?
1000 Currently, we join subqueries to outer queries by sequentially
1001 scanning the result of the subquery for each row of the outer query.
1002 If the subquery returns only a few rows and the outer query returns
1003 many rows, IN is fastest. To speed up other queries, replace IN with
1007 WHERE col IN (SELECT subcol FROM subtab);
1012 WHERE EXISTS (SELECT subcol FROM subtab WHERE subcol = col);
1014 For this to be fast, subcol should be an indexed column. We hope to
1015 fix this limitation in a future release.
1017 4.23) How do I perform an outer join?
1019 PostgreSQL supports outer joins using the SQL standard syntax. Here
1022 FROM t1 LEFT OUTER JOIN t2 ON (t1.col = t2.col);
1026 FROM t1 LEFT OUTER JOIN t2 USING (col);
1028 These identical queries join t1.col to t2.col, and also return any
1029 unjoined rows in t1 (those with no match in t2). A RIGHT join would
1030 add unjoined rows of t2. A FULL join would return the matched rows
1031 plus all unjoined rows from t1 and t2. The word OUTER is optional and
1032 is assumed in LEFT, RIGHT, and FULL joins. Ordinary joins are called
1035 In previous releases, outer joins can be simulated using UNION and NOT
1036 IN. For example, when joining tab1 and tab2, the following query does
1037 an outer join of the two tables:
1038 SELECT tab1.col1, tab2.col2
1040 WHERE tab1.col1 = tab2.col1
1042 SELECT tab1.col1, NULL
1044 WHERE tab1.col1 NOT IN (SELECT tab2.col1 FROM tab2)
1047 4.24) How do I perform queries using multiple databases?
1049 There is no way to query a database other than the current one.
1050 Because PostgreSQL loads database-specific system catalogs, it is
1051 uncertain how a cross-database query should even behave.
1053 contrib/dblink allows cross-database queries using function calls. Of
1054 course, a client can make simultaneous connections to different
1055 databases and merge the results on the client side.
1057 4.25) How do I return multiple rows or columns from a function?
1059 You can return result sets from PL/pgSQL functions using refcursors.
1060 See http://www.PostgreSQL.org/idocs/index.php?plpgsql-cursors.html,
1063 4.26) Why can't I reliably create/drop temporary tables in PL/PgSQL
1066 PL/PgSQL caches function contents, and an unfortunate side effect is
1067 that if a PL/PgSQL function accesses a temporary table, and that table
1068 is later dropped and recreated, and the function called again, the
1069 function will fail because the cached function contents still point to
1070 the old temporary table. The solution is to use EXECUTE for temporary
1071 table access in PL/PgSQL. This will cause the query to be reparsed
1074 4.27) What replication options are available?
1076 There are several master/slave replication options available. These
1077 allow only the master to make database changes and the slave can only
1078 do database reads. The bottom of
1079 http://gborg.PostgreSQL.org/genpage?replication_research lists them. A
1080 multi-master replication solution is being worked on at
1081 http://gborg.PostgreSQL.org/project/pgreplication/projdisplay.php.
1083 4.28) What encryption options are available?
1085 * contrib/pgcrypto contains many encryption functions for use in SQL
1087 * The only way to encrypt transmission from the client to the server
1088 is by using hostssl in pg_hba.conf.
1089 * Database user passwords are automatically encrypted when stored in
1090 version 7.3. In previous versions, you must enable the option
1091 PASSWORD_ENCRYPTION in postgresql.conf.
1092 * The server can run using an encrypted file system.
1093 _________________________________________________________________
1095 Extending PostgreSQL
1097 5.1) I wrote a user-defined function. When I run it in psql, why does it
1100 The problem could be a number of things. Try testing your user-defined
1101 function in a stand-alone test program first.
1103 5.2) How can I contribute some nifty new types and functions to PostgreSQL?
1105 Send your extensions to the pgsql-hackers mailing list, and they will
1106 eventually end up in the contrib/ subdirectory.
1108 5.3) How do I write a C function to return a tuple?
1110 In versions of PostgreSQL beginning with 7.3, table-returning
1111 functions are fully supported in C, PL/PgSQL, and SQL. See the
1112 Programmer's Guide for more information. An example of a
1113 table-returning function defined in C can be found in
1116 5.4) I have changed a source file. Why does the recompile not see the
1119 The Makefiles do not have the proper dependencies for include files.
1120 You have to do a make clean and then another make. If you are using
1121 GCC you can use the --enable-depend option of configure to have the
1122 compiler compute the dependencies automatically.