3 Current maintainer: Bruce Momjian (bruce@momjian.us)
4 Last updated: Mon Jun 23 18:23:48 EDT 2008
6 The most recent version of this document can be viewed at
7 http://www.postgresql.org/docs/faqs.TODO.html.
9 #A hyphen, "-", marks changes that will appear in the upcoming 8.4 release.#
10 #A percent sign, "%", marks items that are easier to implement.#
12 This list contains all known PostgreSQL bugs and feature requests. If
13 you would like to work on an item, please read the Developer's FAQ
14 first. There is also a developer's wiki at
15 http://developer.postgresql.org.
21 * -Allow administrators to safely terminate individual sessions either
22 via an SQL function or SIGTERM
23 * Check for unreferenced table files created by transactions that were
24 in-progress when the server terminated abruptly
26 http://archives.postgresql.org/pgsql-patches/2006-06/msg00096.php
28 * Set proper permissions on non-system schemas during db creation
30 Currently all schemas are owned by the super-user because they are copied
31 from the template1 database. However, since all objects are inherited
32 from the template database, it is not clear that setting schemas to the db
35 * -Add function to report the time of the most recent server reload
36 * Allow log_min_messages to be specified on a per-module basis
38 This would allow administrators to see more detailed information from
39 specific sections of the backend, e.g. checkpoints, autovacuum, etc.
40 Another idea is to allow separate configuration files for each module,
41 or allow arbitrary SET commands to be passed to them.
43 * Simplify ability to create partitioned tables
45 This would allow creation of partitioned tables without requiring
46 creation of triggers or rules for INSERT/UPDATE/DELETE, and constraints
47 for rapid partition selection. Options could include range and hash
50 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00375.php
51 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00151.php
52 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00028.php
53 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00248.php
54 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00387.php
55 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00413.php
57 * Allow auto-selection of partitioned tables for min/max() operations
58 * Allow more complex user/database default GUC settings
60 Currently ALTER USER and ALTER DATABASE support per-user and
61 per-database defaults. Consider adding per-user-and-database
62 defaults so things like search_path can be defaulted for a
63 specific user connecting to a specific database.
65 * Allow custom variables to appear in pg_settings()
66 * Allow custom variable classes that can restrict who can set the values
68 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00911.php
70 * Implement the SQL standard mechanism whereby REVOKE ROLE revokes only
71 the privilege granted by the invoking role, and not those granted
74 http://archives.postgresql.org/pgsql-bugs/2007-05/msg00010.php
76 * Allow SSL authentication/encryption over unix domain sockets
78 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00924.php
80 * Allow SSL key file permission checks to be optionally disabled when
81 sharing SSL keys with other applications
83 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00069.php
85 * Allow client certificate names to be checked against the client
88 This is already implemented in
89 libpq/fe-secure.c::verify_peer_name_matches_certificate() but the code
94 o Allow pg_hba.conf to specify host names along with IP addresses
96 Host name lookup could occur when the postmaster reads the
97 pg_hba.conf file, or when the backend starts. Another
98 solution would be to reverse lookup the connection IP and
99 check that hostname against the host names in pg_hba.conf.
100 We could also then check that the host name maps to the IP
102 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00569.php
104 o %Allow postgresql.conf file values to be changed via an SQL
105 API, perhaps using SET GLOBAL
106 o Allow the server to be stopped/restarted via an SQL API
107 o Issue a warning if a change-on-restart-only postgresql.conf value
108 is modified and the server config files are reloaded
109 o Consider normalizing fractions in postgresql.conf, perhaps
112 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00550.php
114 o Allow Kerberos to disable stripping of realms so we can
115 check the username@realm against multiple realms
117 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00009.php
119 o Add functions to syntax check configuration files
124 o Allow a database in tablespace t1 with tables created in
125 tablespace t2 to be used as a template for a new database created
126 with default tablespace t2
128 Currently all objects in the default database tablespace must
129 have default tablespace specifications. This is because new
130 databases are created by copying directories. If you mix default
131 tablespace tables and tablespace-specified tables in the same
132 directory, creating a new database from such a mixed directory
133 would create a new database with tables that had incorrect
134 explicit tablespaces. To fix this would require modifying
135 pg_class in the newly copied database, which we don't currently
138 o Allow reporting of which objects are in which tablespaces
140 This item is difficult because a tablespace can contain objects
141 from multiple databases. There is a server-side function that
142 returns the databases which use a specific tablespace, so this
143 requires a tool that will call that function and connect to each
144 database to find the objects in each database for that tablespace.
146 o Allow WAL replay of CREATE TABLESPACE to work when the directory
147 structure on the recovery computer is different from the original
149 o Allow per-tablespace quotas
152 * Statistics Collector
154 o Allow statistics collector information to be pulled from the collector
155 process directly, rather than requiring the collector to write a
156 filesystem file twice a second?
157 o Reduce file system activity overhead of statistics file pgstat.stat
159 http://archives.postgresql.org/pgsql-general/2007-12/msg00106.php
161 o Allow statistics last vacuum/analyze execution times to be displayed
162 without requiring stats_row_level to be enabled
164 http://archives.postgresql.org/pgsql-docs/2007-04/msg00028.php
166 o Clear table counters on TRUNCATE
168 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00169.php
171 * Point-In-Time Recovery (PITR)
173 o Allow a warm standby system to also allow read-only statements
175 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00050.php
177 o %Create dump tool for write-ahead logs for use in determining
178 transaction id for point-in-time recovery
180 This is useful for checking PITR recovery.
182 o Allow recovery.conf to support the same syntax as
183 postgresql.conf, including quoting
185 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00497.php
187 o Fix server restart problem when the server was shutdown during
190 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00800.php
192 o Recreate pg_xlog/archive_status/ if it doesn't exist after
193 restoring from a PITR backup
195 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00487.php
201 * Change NUMERIC to enforce the maximum precision
202 * Reduce storage space for small NUMERICs
204 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01331.php
205 http://archives.postgresql.org/pgsql-patches/2007-02/msg00505.php
206 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00715.php
208 * Fix data types where equality comparison isn't intuitive, e.g. box
209 * Add support for public SYNONYMs
211 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00519.php
213 * Fix CREATE CAST on DOMAINs
215 http://archives.postgresql.org/pgsql-hackers/2006-05/msg00072.php
216 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01681.php
218 * Allow domains to be cast
220 http://archives.postgresql.org/pgsql-hackers/2003-06/msg01206.php
221 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00289.php
223 * Add support for SQL-standard GENERATED/IDENTITY columns
225 http://archives.postgresql.org/pgsql-hackers/2006-07/msg00543.php
226 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00038.php
227 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00344.php
228 http://archives.postgresql.org/pgsql-patches/2007-05/msg00076.php
229 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00604.php
231 * Improve XML support
233 http://developer.postgresql.org/index.php/XML_Support
235 * Consider placing all sequences in a single table, or create a system
238 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
240 * Allow the UUID type to accept non-standard formats
242 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01214.php
244 * Consider a special data type for regular expressions
246 http://archives.postgresql.org/pgsql-hackers/2007-08/msg01067.php
248 * Reduce BIT data type overhead using short varlena headers
250 http://archives.postgresql.org/pgsql-general/2007-12/msg00273.php
252 * Allow xml arrays to be cast to other data types
254 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00981.php
255 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00231.php
256 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00471.php
258 * Simplify integer cross-data-type operators
260 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00189.php
262 * Allow XML to accept more liberal DOCTYPE specifications
264 http://archives.postgresql.org/pgsql-general/2008-02/msg00347.php
266 * Allow adding/renaming/removing enumerated values to an existing
269 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01718.php
273 o Allow infinite dates and intervals just like infinite timestamps
274 o Merge hardwired timezone names with the TZ database; allow either
275 kind everywhere a TZ name is currently taken
276 o Allow TIMESTAMP WITH TIME ZONE to store the original timezone
277 information, either zone name or offset from UTC
279 If the TIMESTAMP value is stored with a time zone name, interval
280 computations should adjust based on the time zone rules.
281 http://archives.postgresql.org/pgsql-hackers/2004-10/msg00705.php
283 o Fix SELECT '0.01 years'::interval, '0.01 months'::interval
284 o Add a GUC variable to allow output of interval values in ISO8601
286 o Have timestamp subtraction not call justify_hours()?
288 http://archives.postgresql.org/pgsql-sql/2006-10/msg00059.php
290 o Improve timestamptz subtraction to be DST-aware
292 Currently subtracting one date from another that crosses a
293 daylight savings time adjustment can return '1 day 1 hour', but
294 adding that back to the first date returns a time one hour in
295 the future. This is caused by the adjustment of '25 hours' to
296 '1 day 1 hour', and '1 day' is the same time the next day, even
297 if daylight savings adjustments are involved.
299 o Fix interval display to support values exceeding 2^31 hours
300 o Add overflow checking to timestamp and interval arithmetic
301 o Extend timezone code to allow 64-bit values so we can
302 represent years beyond 2038
304 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01363.php
306 o -Use LC_TIME for localized weekday/month names, rather than
309 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00390.php
311 o Add ISO INTERVAL handling
313 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00250.php
314 http://archives.postgresql.org/pgsql-bugs/2006-04/msg00248.php
316 o Support ISO INTERVAL syntax if units cannot be determined from
317 the string, and are supplied after the string
319 The SQL standard states that the units after the string
320 specify the units of the string, e.g. INTERVAL '2' MINUTE
321 should return '00:02:00'. The current behavior has the units
322 restrict the interval value to the specified unit or unit
323 range, INTERVAL '70' SECOND returns '00:00:10'.
325 For syntax that isn't uniquely ISO or PG syntax, like '1' or
326 '1:30', treat as ISO if there is a range specification clause,
327 and as PG if there no clause is present, e.g. interpret '1:30'
328 MINUTE TO SECOND as '1 minute 30 seconds', and interpret
329 '1:30' as '1 hour, 30 minutes'.
331 This makes common cases like SELECT INTERVAL '1' MONTH
332 SQL-standard results. The SQL standard supports a limited
333 number of unit combinations and doesn't support unit names in
334 the string. The PostgreSQL syntax is more flexible in the
335 range of units supported, e.g. PostgreSQL supports '1 year 1
336 hour', while the SQL standard does not.
338 o Add support for year-month syntax, INTERVAL '50-6' YEAR
340 o Interpret INTERVAL '1 year' MONTH as CAST (INTERVAL '1
341 year' AS INTERVAL MONTH), and this should return '12 months'
342 o Round or truncate values to the requested precision, e.g.
343 INTERVAL '11 months' AS YEAR should return one or zero
344 o Support precision, CREATE TABLE foo (a INTERVAL MONTH(3))
349 o Delay resolution of array expression's data type so assignment
350 coercion can be performed on empty array expressions
351 o Add support for arrays of domains
353 http://archives.postgresql.org/pgsql-patches/2007-05/msg00114.php
355 o Allow single-byte header storage for array elements
360 o Improve vacuum of large objects, like contrib/vacuumlo?
361 o Add security checking for large objects
362 o Auto-delete large objects when referencing row is deleted
364 contrib/lo offers this functionality.
366 o Allow read/write into TOAST values like large objects
368 This requires the TOAST column to be stored EXTERNAL.
370 o Add API for 64-bit large object access
372 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00781.php
376 * Add locale-aware MONEY type, and support multiple currencies
378 http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php
379 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php
381 * MONEY dumps in a locale-specific format making it difficult to
382 restore to a system with a different locale
383 * Allow MONEY to be easily cast to/from other numeric data types
387 o Allow dictionaries to change the token that is passed on to
390 http://archives.postgresql.org/pgsql-patches/2007-11/msg00081.php
392 o Consider a function-based API for '@@' searches
394 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00511.php
396 o Improve text search error messages
398 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00966.php
399 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01146.php
401 o Consider changing error to warning for strings larger than one
404 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00190.php
405 http://archives.postgresql.org/pgsql-patches/2008-03/msg00062.php
412 * Allow INET subnet tests using non-constants to be indexed
413 * Allow to_date() and to_timestamp() accept localized month names
414 * Fix to_date()-related functions to consistently issue errors
416 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00915.php
418 * Add missing parameter handling in to_char()
420 http://archives.postgresql.org/pgsql-hackers/2005-12/msg00948.php
422 * Allow substring/replace() to get/set bit values
423 * Allow to_char() on interval values to accumulate the highest unit
426 Some special format flag would be required to request such
427 accumulation. Such functionality could also be added to EXTRACT.
428 Prevent accumulation that crosses the month/day boundary because of
429 the uneven number of days in a month.
431 o to_char(INTERVAL '1 hour 5 minutes', 'MI') => 65
432 o to_char(INTERVAL '43 hours 20 minutes', 'MI' ) => 2600
433 o to_char(INTERVAL '43 hours 20 minutes', 'WK:DD:HR:MI') => 0:1:19:20
434 o to_char(INTERVAL '3 years 5 months','MM') => 41
436 * Implement inlining of set-returning functions defined in SQL
437 * Allow SQL-language functions to return results from RETURNING queries
439 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00665.php
441 * Allow SQL-language functions to reference parameters by parameter name
443 Currently SQL-language functions can only refer to dollar parameters,
446 * Add SPI_gettypmod() to return the typemod for a TupleDesc
447 * Enforce typmod for function inputs, function results and parameters for
448 spi_prepare'd statements called from PLs
450 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01403.php
452 * Allow holdable cursors in SPI
453 * Tighten function permission checks
455 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00568.php
457 * Fix IS OF so it matches the ISO specification, and add documentation
459 http://archives.postgresql.org/pgsql-patches/2003-08/msg00060.php
460 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00060.php
462 * Add missing operators for geometric data types
464 Some geometric types do not have the full suite of geometric operators,
467 * Implement Boyer-Moore searching in strpos()
469 http://archives.postgresql.org/pgsql-patches/2007-08/msg00012.php
471 * Prevent malicious functions from being executed with the permissions
472 of unsuspecting users
474 Index functions are safe, so VACUUM and ANALYZE are safe too.
475 Triggers, CHECK and DEFAULT expressions, and rules are still vulnerable.
476 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00268.php
478 * Reduce memory usage of aggregates in set returning functions
480 http://archives.postgresql.org/pgsql-performance/2008-01/msg00031.php
482 * Add temporal versions of generate_series()
484 http://archives.postgresql.org/pgsql-hackers/2007-04/msg01180.php
486 * Add array_accum() and array_to_set() functions for arrays
488 The standards specify array_agg() and UNNEST.
489 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00464.php
491 * Fix /contrib/ltree operator
493 http://archives.postgresql.org/pgsql-bugs/2007-11/msg00044.php
495 * Fix inconsistent precedence of =, >, and < compared to <>, >=, and <=
497 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00145.php
499 * Fix regular expression bug when using complex back-references
501 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00000.php
503 * Have /contrib/dblink reuse unnamed connections
505 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00895.php
507 * Add SQL-standard array_agg() and unnest() array functions
509 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01017.php
511 * Allow calling of a procedure outside a SELECT that can control the
514 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01375.php
518 Multi-Language Support
519 ======================
521 * Add NCHAR (as distinguished from ordinary varchar),
522 * Allow locale to be set at database creation
524 Currently locale can only be set during initdb. No global tables have
525 locale-aware columns. However, the database template used during
526 database creation might have locale-aware indexes. The indexes would
527 need to be reindexed to match the new locale.
529 * Allow encoding on a per-column basis optionally using the ICU library;
532 Right now only one encoding is allowed per database.
534 http://archives.postgresql.org/pgsql-hackers/2005-03/msg00932.php
535 http://archives.postgresql.org/pgsql-patches/2005-08/msg00039.php
536 http://archives.postgresql.org/pgsql-patches/2005-08/msg00309.php
537 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00110.php
538 http://archives.postgresql.org/pgsql-patches/2005-09/msg00020.php
539 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01121.php
540 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00767.php
541 http://archives.postgresql.org/pgsql-patches/2006-03/msg00233.php
542 http://archives.postgresql.org/pgsql-hackers/2006-09/msg00662.php
543 http://wiki.postgresql.org/wiki/Todo:Collate
544 http://wiki.postgresql.org/wiki/Todo:ICU
546 * Support multiple simultaneous character sets, per SQL92
547 * Improve UTF8 combined character handling?
548 * Add octet_length_server() and octet_length_client()
549 * Make octet_length_client() the same as octet_length()?
550 * Fix problems with wrong runtime encoding conversion for NLS message files
551 * Add URL to more complete multi-byte regression tests
553 http://archives.postgresql.org/pgsql-hackers/2005-07/msg00272.php
555 * Fix ILIKE and regular expressions to handle case insensitivity
556 properly in multibyte encodings
558 http://archives.postgresql.org/pgsql-bugs/2005-10/msg00001.php
559 http://archives.postgresql.org/pgsql-patches/2005-11/msg00173.php
561 * Set client encoding based on the client operating system encoding
563 Currently client_encoding is set in postgresql.conf, which
564 defaults to the server encoding.
565 http://archives.postgresql.org/pgsql-hackers/2006-08/msg01696.php
567 * Change memory allocation for multi-byte functions so memory is
568 allocated inside conversion functions
570 Currently we preallocate memory based on worst-case usage.
577 * Automatically create rules on views so they are updateable, per SQL99
579 We can only auto-create rules for simple views. For more complex
580 cases users will still have to write rules manually.
582 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00586.php
583 http://archives.postgresql.org/pgsql-patches/2006-08/msg00255.php
585 * Add the functionality for WITH CHECK OPTION clause of CREATE VIEW
586 * Allow VIEW/RULE recompilation when the underlying tables change
588 Another issue is whether underlying table changes should be reflected
589 in the view, e.g. should SELECT * show additional columns if they
590 are added after the view is created.
592 * Make it possible to use RETURNING together with conditional DO INSTEAD
593 rules, such as for partitioning setups
595 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00577.php
597 * Add the ability to automatically create materialized views
599 Right now materialized views require the user to create triggers on the
600 main table to keep the summary table current. SQL syntax should be able
601 to manager the triggers and summary table automatically. A more
602 sophisticated implementation would automatically retrieve from the
603 summary table when the main table is referenced, if possible.
610 * Add CORRESPONDING BY to UNION/INTERSECT/EXCEPT
611 * Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY
612 * %Allow SET CONSTRAINTS to be qualified by schema/table name
613 * %Add a separate TRUNCATE permission
615 Currently only the owner can TRUNCATE a table because triggers are not
616 called, and the table is locked in exclusive mode.
618 * Allow PREPARE of cursors
619 * Allow finer control over the caching of prepared query plans
621 Currently queries prepared via the libpq API are planned on first
622 execute using the supplied parameters --- allow SQL PREPARE to do the
623 same. Also, allow control over replanning prepared queries either
624 manually or automatically when statistics for execute parameters
625 differ dramatically from those used during planning.
627 * Improve logging of prepared transactions recovered during startup
629 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00092.php
631 * Improve failure message when DROP DATABASE is used on a database that
632 has prepared transactions
633 * Allow prepared transactions with temporary tables created and dropped
634 in the same transaction, and when an ON COMMIT DELETE ROWS temporary
637 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00047.php
639 * Add a GUC variable to warn about non-standard SQL usage in queries
640 * Add SQL-standard MERGE/REPLACE/UPSERT command
642 MERGE is typically used to merge two tables. REPLACE or UPSERT
643 command does UPDATE, or on failure, INSERT. This is similar to UPDATE,
644 then for unmatched rows, INSERT. Whether concurrent access allows
645 modifications which could cause row loss is implementation independent.
646 To implement this cleanly requires that the table have a unique index
647 so duplicate checking can be easily performed. It is possible to do it
648 without a unique index if we require the user to LOCK the table before
651 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00501.php
652 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00536.php
653 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01157.php
654 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01475.php
656 * Add NOVICE output level for helpful messages like automatic sequence/index
658 * Add GUC to issue notice about statements that use unjoined tables
659 * Allow EXPLAIN to identify tables that were skipped because of
661 * Allow EXPLAIN output to be more easily processed by scripts, perhaps XML
662 * Enable standard_conforming_strings
663 * Make standard_conforming_strings the default in 8.5?
665 When this is done, backslash-quote should be prohibited in non-E''
666 strings because of possible confusion over how such strings treat
667 backslashes. Basically, '' is always safe for a literal single
668 quote, while \' might or might not be based on the backslash
671 * Simplify dropping roles that have objects in several databases
672 * Allow COMMENT ON to accept an expression rather than just a string
673 * Allow the count returned by SELECT, etc to be represented as an int64
674 to allow a higher range of values
675 * Add SQL99 WITH clause to SELECT
676 * Add SQL:2003 WITH RECURSIVE (hierarchical) queries to SELECT
678 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01375.php
679 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00642.php
680 http://archives.postgresql.org/pgsql-patches/2007-03/msg00139.php
681 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01334.php
682 http://archives.postgresql.org/pgsql-patches/2008-01/msg00105.php
683 http://archives.postgresql.org/pgsql-patches/2008-03/msg00327.php
685 * Add DEFAULT .. AS OWNER so permission checks are done as the table
688 This would be useful for SERIAL nextval() calls and CHECK constraints.
690 * Allow DISTINCT to work in multiple-argument aggregate calls
691 * Add column to pg_stat_activity that shows the progress of long-running
692 commands like CREATE INDEX and VACUUM
694 http://archives.postgresql.org/pgsql-patches/2008-04/msg00203.php
696 * Implement SQL:2003 window functions
697 * Allow INSERT/UPDATE ... RETURNING inside a SELECT 'FROM' clause or
700 http://archives.postgresql.org/pgsql-general/2006-09/msg00803.php
701 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00693.php
702 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00124.php
704 * Increase locking when DROPing objects so dependent objects cannot
705 get dropped while the DROP operation is happening
707 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00937.php
709 * -Allow AS in "SELECT col AS label" to be optional in certain cases
711 * Allow INSERT ... DELETE ... RETURNING, namely allow the DELETE ...
712 RETURNING to supply values to the INSERT
713 http://archives.postgresql.org/pgsql-hackers/2008-02/thrd2.php#00979
715 * Add comments on system tables/columns using the information in
718 Ideally the information would be pulled from the SGML file
721 * Improve reporting of UNION type mismatches
723 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00944.php
724 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00597.php
729 o Allow CREATE TABLE AS to determine column lengths for complex
730 expressions like SELECT col1 || col2
732 o Have WITH CONSTRAINTS also create constraint indexes
734 http://archives.postgresql.org/pgsql-patches/2007-04/msg00149.php
736 o Have CONSTRAINT cname NOT NULL record the contraint name
738 Right now pg_attribute.attnotnull records the NOT NULL status
739 of the column, but does not record the contraint name
741 o Prevent concurrent CREATE TABLE table1 from sometimes returning
742 a cryptic error message
744 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00169.php
746 o Add CREATE SCHEMA ... LIKE that copies a schema
751 o Allow UPDATE tab SET ROW (col, ...) = (SELECT...)
753 http://archives.postgresql.org/pgsql-hackers/2006-07/msg01306.php
754 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00865.php
755 http://archives.postgresql.org/pgsql-patches/2007-04/msg00315.php
756 http://archives.postgresql.org/pgsql-patches/2008-03/msg00237.php
758 o Research self-referential UPDATEs that see inconsistent row versions
759 in read-committed mode
761 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00507.php
762 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00016.php
764 o Allow GLOBAL temporary tables to exist as empty by default in
767 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00006.php
772 o Have ALTER TABLE RENAME rename SERIAL sequence names
774 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
776 o Have ALTER SEQUENCE RENAME rename the sequence name stored
777 in the sequence table
779 http://archives.postgresql.org/pgsql-bugs/2007-09/msg00092.php
780 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00007.php
781 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
783 o Add ALTER DOMAIN to modify the underlying data type
784 o %Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME
786 http://archives.postgresql.org/pgsql-patches/2006-02/msg00168.php
788 o %Allow ALTER TABLE to change constraint deferrability and actions
789 o Add missing object types for ALTER ... SET SCHEMA
790 o Allow ALTER TABLESPACE to move to different directories
791 o Allow databases to be moved to different tablespaces
792 o Allow moving system tables to other tablespaces, where possible
794 Currently non-global system tables must be in the default database
795 tablespace. Global system tables can never be moved.
797 o -Prevent parent tables from altering or dropping constraints
798 like CHECK that are inherited by child tables unless CASCADE
800 o -Prevent child tables from altering or dropping constraints
801 like CHECK that were inherited from the parent table
802 o Have ALTER INDEX update the name of a constraint using that index
803 o Add ALTER TABLE RENAME CONSTRAINT, update index name also
804 o Allow column display reordering by recording a display,
805 storage, and permanent id for every column?
807 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00782.php
809 o Allow an existing index to be marked as a table's primary key
811 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00500.php
816 o Automatically maintain clustering on a table
818 This might require some background daemon to maintain clustering
819 during periods of low usage. It might also require tables to be only
820 partially filled for easier reorganization. Another idea would
821 be to create a merged heap/index data file so an index lookup would
822 automatically access the heap data too. A third idea would be to
823 store heap rows in hashed groups, perhaps using a user-supplied
825 http://archives.postgresql.org/pgsql-performance/2004-08/msg00349.php
827 o %Add default clustering to system tables
829 To do this, determine the ideal cluster index for each system
830 table and set the cluster setting during initdb.
832 o %Add VERBOSE option to report tables as they are processed,
838 o Allow COPY to report error lines and continue
840 This requires the use of a savepoint before each COPY line is
841 processed, with ROLLBACK on COPY failure.
842 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00572.php
844 o Allow COPY on a newly-created table to skip WAL logging
846 On crash recovery, the table involved in the COPY would
847 be removed or have its heap and index files truncated. One
848 issue is that no other backend should be able to add to
849 the table at the same time, which is something that is
850 currently allowed. This currently is done if the table is
851 created inside the same transaction block as the COPY because
852 no other backends can see the table.
854 o Consider using a ring buffer for COPY FROM
856 http://archives.postgresql.org/pgsql-patches/2008-02/msg00140.php
857 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php
859 o Allow COPY FROM to create index entries in bulk
861 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00811.php
863 o Allow COPY in CSV mode to control whether a quoted zero-length
864 string is treated as NULL
866 Currently this is always treated as a zero-length string,
867 which generates an error when loading into an integer column
868 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00905.php
870 o Impove COPY performance
872 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00954.php
878 o Allow column-level privileges
879 o %Allow GRANT/REVOKE permissions to be applied to all schema objects
882 The proposed syntax is:
883 GRANT SELECT ON ALL TABLES IN public TO phpuser;
884 GRANT SELECT ON NEW TABLES IN public TO phpuser;
886 o Allow GRANT/REVOKE permissions to be inherited by objects based on
889 o Allow SERIAL sequences to inherit permissions from the base table?
894 o Prevent DROP TABLE from dropping a row referenced by its own open
900 o Allow INSERT/UPDATE of the system-generated oid value for a row
901 o In rules, allow VALUES() to contain a mixture of 'old' and 'new'
907 o Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
913 o Allow LISTEN/NOTIFY to store info in memory rather than tables?
915 Currently LISTEN/NOTIFY information is stored in pg_listener.
916 Storing such information in memory would improve performance.
918 o Add optional textual message to NOTIFY
920 This would allow an informational message to be added to the notify
921 message, perhaps indicating the row modified or other custom
924 o Allow multiple identical NOTIFY events to always be communicated
925 to the client, rather than sent as a single notification to the
928 http://archives.postgresql.org/pgsql-general/2008-01/msg00057.php
930 o Allow NOTIFY in rules involving conditionals
931 o Improve LISTEN concurrency
933 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01106.php
937 Referential Integrity
938 =====================
940 * Add MATCH PARTIAL referential integrity
941 * Change foreign key constraint for array -> element to mean element
943 * Fix problem when cascading referential triggers make changes on
944 cascaded tables, seeing the tables in an intermediate state
946 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
947 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
949 * Allow DEFERRABLE and end-of-statement UNIQUE constraints?
951 This would allow UPDATE tab SET col = col + 1 to work if col has
952 a unique index. Currently, uniqueness checks are done while the
953 command is being executed, rather than at the end of the statement
955 http://people.planetpostgresql.org/greg/index.php?/archives/2006/06/10.html
956 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01458.php
958 * Optimize referential integrity checks
960 http://archives.postgresql.org/pgsql-performance/2005-10/msg00458.php
961 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00744.php
964 Server-Side Languages
965 =====================
968 o Fix RENAME to work on variables other than OLD/NEW
970 http://archives.postgresql.org/pgsql-hackers/2002-03/msg00591.php
971 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01615.php
972 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01587.php
974 o Allow function parameters to be passed by name,
975 get_employee_salary(12345 AS emp_id, 2001 AS tax_year)
976 o Allow handling of %TYPE arrays, e.g. tab.col%TYPE[]
977 o Allow listing of record column names, and access to
978 record columns via variables, e.g. columns := r.(*),
981 http://archives.postgresql.org/pgsql-patches/2005-07/msg00458.php
982 http://archives.postgresql.org/pgsql-patches/2006-05/msg00302.php
983 http://archives.postgresql.org/pgsql-patches/2006-06/msg00031.php
985 o Add support for SCROLL cursors
986 o Add support for WITH HOLD cursors
987 o Allow row and record variables to be set to NULL constants,
988 and allow NULL tests on such variables
990 Because a row is not scalar, do not allow assignment
991 from NULL-valued scalars.
993 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00070.php
995 o Review handling of MOVE and FETCH
997 http://archives.postgresql.org/pgsql-patches/2007-04/msg00527.php
999 o Improve logic of determining if an identifier is a a
1000 variable or column name
1002 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00436.php
1004 o Consider keeping seperate cached copies when search_path changes
1006 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01009.php
1008 o -Add CASE capability to language (already in SQL)
1010 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00696.php
1016 o Add table function support to pltcl, plpythonu
1017 o Add support for polymorphic arguments and return types to
1018 languages other than PL/PgSQL
1019 o Add capability to create and call PROCEDURES
1020 o Add support for OUT and INOUT parameters to languages other
1022 o Add PL/PythonU tracebacks
1024 http://archives.postgresql.org/pgsql-patches/2006-02/msg00288.php
1026 o Allow data to be passed in native language formats, rather
1029 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00289.php
1031 o Add ability to obfuscate function bodies
1033 http://archives.postgresql.org/pgsql-patches/2008-01/msg00125.php
1040 * Have pg_ctl look at PGHOST in case it is a socket directory?
1041 * Allow pg_ctl to work properly with configuration files located outside
1042 the PGDATA directory
1044 pg_ctl can not read the pid file because it isn't located in the
1045 config directory but in the PGDATA directory. The solution is to
1046 allow pg_ctl to read and understand postgresql.conf to find the
1047 data_directory value.
1049 * Add a function like pg_get_indexdef() that report more detailed index
1052 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00166.php
1055 o Have psql show current values for a sequence
1056 o Move psql backslash database information into the backend, use
1059 This would allow non-psql clients to pull the same information out
1060 of the database as psql.
1061 http://archives.postgresql.org/pgsql-hackers/2004-01/msg00191.php
1063 o Make psql's \d commands more consistent
1065 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1066 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1068 o Consistently display privilege information for all objects in psql
1069 o Add auto-expanded mode so expanded output is used if the row
1070 length is wider than the screen width.
1072 Consider using auto-expanded mode for backslash commands like \df+.
1074 o Prevent tab completion of SET TRANSACTION from querying the
1075 database and therefore preventing the transaction isolation
1076 level from being set.
1078 Currently SET <tab> causes a database lookup to check all
1079 supported session variables. This query causes problems
1080 because setting the transaction isolation level must be the
1081 first statement of a transaction.
1083 o Add a \set variable to control whether \s displays line numbers
1085 Another option is to add \# which lists line numbers, and
1086 allows command execution.
1088 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00255.php
1090 o Prevent escape string warnings when object names have
1093 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00227.php
1095 o Have \d show foreign keys that reference a table's primary key
1097 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00424.php
1099 o Have \d show child tables that inherit from the specified parent
1100 o -Have \l+ show database size, if permissions allow
1102 Ideally it will not generate an error for invalid permissions
1104 o Include the symbolic SQLSTATE name in verbose error reports
1106 http://archives.postgresql.org/pgsql-general/2007-09/msg00438.php
1108 o -Improve display of enums to show valid enum values
1109 o Add prompt escape to display the client and server versions
1112 * pg_dump / pg_restore
1113 o %Add dumping of comments on index columns and composite type columns
1114 o %Add full object name to the tag field. eg. for operators we need
1115 '=(integer, integer)', instead of just '='.
1116 o Add pg_dumpall custom format dumps?
1117 o Allow selection of individual object(s) of all types, not just
1119 o In a selective dump, allow dumping of an object and all its
1121 o Add options like pg_restore -l and -L to pg_dump
1122 o Stop dumping CASCADE on DROP TYPE commands in clean mode
1123 o Allow pg_dump --clean to drop roles that own objects or have
1125 o Change pg_dump so that a comment on the dumped database is
1126 applied to the loaded database, even if the database has a
1127 different name. This will require new backend syntax, perhaps
1128 COMMENT ON CURRENT DATABASE.
1129 o Remove unnecessary function pointer abstractions in pg_dump source
1131 o Allow pg_dump to utilize multiple CPUs and I/O channels by dumping
1132 multiple objects simultaneously
1134 The difficulty with this is getting multiple dump processes to
1135 produce a single dump output file. It also would require
1136 several sessions to share the same snapshot.
1137 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1139 o Allow pg_restore to utilize multiple CPUs and I/O channels by
1140 restoring multiple objects simultaneously
1142 This might require a pg_restore flag to indicate how many
1143 simultaneous operations should be performed. Only pg_dump's
1144 -Fc format has the necessary dependency information.
1145 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00963.php
1147 o To better utilize resources, allow pg_restore to check foreign
1148 keys simultaneously, where possible
1149 o Allow pg_restore to create all indexes of a table
1150 concurrently, via a single heap scan
1152 This requires a pg_dump -Fc file because that format contains
1153 the required dependency information.
1154 http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php
1156 o Allow pg_restore to load different parts of the COPY data
1158 o Prevent pg_dump/pg_restore from being affected by
1161 Using psql to restore a pg_dump dump is also affected.
1163 o Remove pre-7.3 pg_dump code that assumes pg_depend does not exit
1164 o Allow pre/data/post files when schema and data are dumped
1165 separately, for performance reasons
1167 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1174 Document differences between ecpg and the SQL standard and
1175 information about the Informix-compatibility module.
1177 o Solve cardinality > 1 for input descriptors / variables?
1178 o Add a semantic check level, e.g. check if a table really exists
1179 o fix handling of DB attributes that are arrays
1180 o Use backend PREPARE/EXECUTE facility for ecpg where possible
1182 o Fix nested C comments
1183 o %sqlwarn[6] should be 'W' if the PRECISION or SCALE value specified
1184 o Make SET CONNECTION thread-aware, non-standard?
1185 o Allow multidimensional arrays
1186 o Add internationalized message strings
1187 o Implement COPY FROM STDIN
1191 o Add PQescapeIdentifierConn()
1192 o Prevent PQfnumber() from lowercasing unquoted the column name
1194 PQfnumber() should never have been doing lowercasing, but
1195 historically it has so we need a way to prevent it
1197 o Allow statement results to be automatically batched to the client
1199 Currently all statement results are transferred to the libpq
1200 client before libpq makes the results available to the
1201 application. This feature would allow the application to make
1202 use of the first result rows while the rest are transferred, or
1203 held on the server waiting for them to be requested by libpq.
1204 One complexity is that a statement like SELECT 1/col could error
1205 out mid-way through the result set.
1207 o Consider disallowing multiple queries in PQexec() as an
1208 additional barrier to SQL injection attacks
1210 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00184.php
1212 o Add PQexecf() that allows complex parameter substitution
1214 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01803.php
1216 o Add SQLSTATE severity to PGconn return status
1218 http://archives.postgresql.org/pgsql-interfaces/2007-11/msg00015.php
1224 * Add deferred trigger queue file
1226 Right now all deferred trigger information is stored in backend
1227 memory. This could exhaust memory for very large trigger queues.
1228 This item involves dumping large queues into files, or doing some
1229 kind of join to process all the triggers, or some bulk operation.
1230 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00876.php
1232 * Allow triggers to be disabled in only the current session.
1234 This is currently possible by starting a multi-statement transaction,
1235 modifying the system tables, performing the desired SQL, restoring the
1236 system tables, and committing the transaction. ALTER TABLE ...
1237 TRIGGER requires a table lock so it is not ideal for this usage.
1239 * With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY
1241 If the dump is known to be valid, allow foreign keys to be added
1242 without revalidating the data.
1244 * Allow statement-level triggers to access modified rows
1245 * Support triggers on columns
1247 http://archives.postgresql.org/pgsql-patches/2005-07/msg00107.php
1249 * Allow AFTER triggers on system tables
1251 System tables are modified in many places in the backend without going
1252 through the executor and therefore not causing triggers to fire. To
1253 complete this item, the functions that modify system tables will have
1256 * Tighten trigger permission checks
1258 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php
1260 * Allow BEFORE INSERT triggers on views
1262 http://archives.postgresql.org/pgsql-general/2007-02/msg01466.php
1264 * -Add ability to trigger on TRUNCATE
1265 * Add database and transaction-level triggers
1267 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00451.php
1274 * Add UNIQUE capability to non-btree indexes
1275 * Prevent index uniqueness checks when UPDATE does not modify the column
1277 Uniqueness (index) checks are done when updating a column even if the
1278 column is not modified by the UPDATE.
1280 * Allow the creation of on-disk bitmap indexes which can be quickly
1281 combined with other bitmap indexes
1283 Such indexes could be more compact if there are only a few distinct values.
1284 Such indexes can also be compressed. Keeping such indexes updated can be
1287 http://archives.postgresql.org/pgsql-patches/2005-07/msg00512.php
1288 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01107.php
1289 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00265.php
1290 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01214.php
1291 http://archives.postgresql.org/pgsql-patches/2007-05/msg00013.php
1292 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00741.php
1294 * Allow accurate statistics to be collected on indexes with more than
1295 one column or expression indexes, perhaps using per-index statistics
1297 http://archives.postgresql.org/pgsql-performance/2006-10/msg00222.php
1298 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01131.php
1300 * Consider increasing the default and maximum number of statistics targets,
1301 and reduce statistics target overhead
1303 Also consider having a larger statistics target for indexed columns
1304 and expression indexes.
1305 http://archives.postgresql.org/pgsql-general/2007-05/msg01228.php
1306 http://archives.postgresql.org/pgsql-general/2007-06/msg00542.php
1307 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01066.php
1308 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00188.php
1310 * Consider smaller indexes that record a range of values per heap page,
1311 rather than having one index entry for every heap row
1313 This is useful if the heap is clustered by the indexed values.
1314 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00341.php
1315 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01264.php
1316 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00465.php
1317 http://archives.postgresql.org/pgsql-patches/2007-03/msg00163.php
1318 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00014.php
1319 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00487.php
1320 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01589.php
1322 * Add REINDEX CONCURRENTLY, like CREATE INDEX CONCURRENTLY
1324 This is difficult because you must upgrade to an exclusive table lock
1325 to replace the existing index file. CREATE INDEX CONCURRENTLY does not
1326 have this complication. This would allow index compaction without
1328 http://archives.postgresql.org/pgsql-performance/2007-08/msg00289.php
1330 * Allow multiple indexes to be created concurrently, ideally via a
1331 single heap scan, and have pg_restore use it
1333 * Consider sorting entries before inserting into btree index
1335 http://archives.postgresql.org/pgsql-general/2008-01/msg01010.php
1337 * Allow index scans to return matching index keys, not just the matching
1340 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01657.php
1346 o Allow inherited tables to inherit indexes, UNIQUE constraints,
1347 and primary/foreign keys
1348 o Honor UNIQUE INDEX on base column in INSERTs/UPDATEs
1349 on inherited table, e.g. INSERT INTO inherit_table
1350 (unique_index_col) VALUES (dup) should fail
1352 The main difficulty with this item is the problem of
1353 creating an index that can span multiple tables.
1355 o Allow SELECT ... FOR UPDATE on inherited tables
1356 o Require all CHECK constraints to be inherited
1358 http://archives.postgresql.org/pgsql-bugs/2007-04/msg00026.php
1360 o Add checks to prevent a CREATE RULE views on inherited tables
1362 http://archives.postgresql.org/pgsql-general/2008-02/msg01420.php
1363 http://archives.postgresql.org/pgsql-general/2008-03/msg00077.php
1368 o Add more GIST index support for geometric data types
1369 o Allow GIST indexes to create certain complex index types, like
1370 digital trees (see Aoki)
1375 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00051.php
1377 o Pack hash index buckets onto disk pages more efficiently
1379 Currently only one hash bucket can be stored on a page. Ideally
1380 several hash buckets could be stored on a single page and greater
1381 granularity used for the hash algorithm.
1383 http://archives.postgresql.org/pgsql-hackers/2004-06/msg00168.php
1385 o Consider sorting hash buckets so entries can be found using a
1386 binary search, rather than a linear scan
1387 o In hash indexes, consider storing the hash value with or instead
1389 o Add WAL logging for crash recovery
1390 o Allow multi-column hash indexes
1391 o -During index creation, pre-sort the tuples to improve build speed
1395 * Consider using hash buckets to do DISTINCT, rather than sorting
1397 This would be beneficial when there are few distinct values. This is
1398 already used by GROUP BY.
1400 * Consider whether duplicate keys should be sorted by block/offset
1402 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00558.php
1404 * -Avoid tuple some tuple copying in sort routines
1405 * Consider being smarter about memory and external files used during
1408 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01101.php
1409 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00045.php
1411 * Consider detoasting keys before sorting
1417 * Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
1419 Ideally this requires a separate test program that can be run
1420 at initdb time or optionally later. Consider O_SYNC when
1423 * Add program to test if fsync has a delay compared to non-fsync
1424 * Consider sorting writes during checkpoint
1426 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00541.php
1435 We could use a fixed row count and a +/- count to follow MVCC
1436 visibility rules, or a single cached value could be used and
1437 invalidated if anyone modifies the table. Another idea is to
1438 get a count directly from a unique index, but for this to be
1439 faster than a sequential scan it must avoid access to the heap
1440 to obtain tuple visibility information.
1442 * Provide a way to calculate an "estimated COUNT(*)"
1444 Perhaps by using the optimizer's cardinality estimates or random
1447 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00943.php
1449 * Allow data to be pulled directly from indexes
1451 Currently indexes do not have enough tuple visibility information
1452 to allow data to be pulled from the index without also accessing
1453 the heap. One way to allow this is to set a bit on index tuples
1454 to indicate if a tuple is currently visible to all transactions
1455 when the first valid heap lookup happens. This bit would have to
1456 be cleared when a heap tuple is expired.
1458 Another idea is to maintain a bitmap of heap pages where all rows
1459 are visible to all backends, and allow index lookups to reference
1460 that bitmap to avoid heap lookups, perhaps the same bitmap we might
1461 add someday to determine which heap pages need vacuuming. Frequently
1462 accessed bitmaps would have to be stored in shared memory. One 8k
1463 page of bitmaps could track 512MB of heap pages.
1465 A third idea would be for a heap scan to check if all rows are visible
1466 and if so set a per-table flag which can be checked by index scans.
1467 Any change to the table would have to clear the flag. To detect
1468 changes during the heap scan a counter could be set at the start and
1469 checked at the end --- if it is the same, the table has not been
1470 modified --- any table change would increment the counter.
1472 http://archives.postgresql.org/pgsql-patches/2007-10/msg00166.php
1473 http://archives.postgresql.org/pgsql-patches/2008-01/msg00049.php
1475 * Consider automatic caching of statements at various levels:
1478 o Query execute plan
1480 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00823.php
1482 * Consider increasing internal areas when shared buffers is increased
1484 http://archives.postgresql.org/pgsql-hackers/2005-10/msg01419.php
1486 * Consider decreasing the amount of memory used by PrivateRefCount
1488 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00797.php
1489 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00752.php
1491 * Consider allowing higher priority queries to have referenced buffer
1492 cache pages stay in memory longer
1494 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00562.php
1501 * Improve speed with indexes
1503 For large table adjustments during VACUUM FULL, it is faster to cluster
1504 or reindex rather than update the index. Also, index updates can bloat
1507 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00024.php
1508 http://archives.postgresql.org/pgsql-performance/2007-05/msg00296.php
1509 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00307.php
1511 * Auto-fill the free space map by scanning the buffer cache or by
1512 checking pages written by the background writer
1514 http://archives.postgresql.org/pgsql-hackers/2006-02/msg01125.php
1515 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00011.php
1517 * Create a bitmap of pages that need vacuuming
1519 Instead of sequentially scanning the entire table, have the background
1520 writer or some other process record pages that have expired rows, then
1521 VACUUM can look at just those pages rather than the entire table. In
1522 the event of a system crash, the bitmap would probably be invalidated.
1523 One complexity is that index entries still have to be vacuumed, and
1524 doing this without an index scan (by using the heap values to find the
1525 index entry) might be slow and unreliable, especially for user-defined
1528 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01188.php
1529 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00121.php
1530 http://archives.postgresql.org/pgsql-patches/2007-03/msg00508.php
1531 http://archives.postgresql.org/pgsql-patches/2007-04/msg00347.php
1532 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00156.php
1533 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00546.php
1534 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00416.php
1536 * Allow FSM to return free space toward the beginning of the heap file,
1537 in hopes that empty pages at the end can be truncated by VACUUM
1538 * Allow FSM page return free space based on table clustering, to assist
1539 in maintaining clustering?
1540 * -Improve dead row detection during multi-statement transactions usage
1542 http://archives.postgresql.org/pgsql-patches/2007-03/msg00358.php
1544 * Consider a more compact data representation for dead tuples
1546 http://archives.postgresql.org/pgsql-patches/2007-05/msg00143.php
1551 o %Issue log message to suggest VACUUM FULL if a table is nearly
1553 o Improve control of auto-vacuum
1555 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00876.php
1557 o Prevent long-lived temporary tables from causing frozen-xid
1558 advancement starvation
1560 The problem is that autovacuum cannot vacuum them to set frozen xids;
1561 only the session that created them can do that.
1562 http://archives.postgresql.org/pgsql-general/2007-06/msg01645.php
1564 o Store per-table autovacuum settings in pg_class.reloptions.
1566 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01440.php
1567 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00724.php
1569 o Prevent autovacuum from running if an old transaction is still
1570 running from the last vacuum
1572 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00899.php
1579 * Fix priority ordering of read and write light-weight locks (Neil)
1581 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00893.php
1582 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00905.php
1584 * Fix problem when multiple subtransactions of the same outer transaction
1585 hold different types of locks, and one subtransaction aborts
1587 http://archives.postgresql.org/pgsql-hackers/2006-11/msg01011.php
1588 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00001.php
1589 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00435.php
1590 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00773.php
1592 * Allow UPDATEs on only non-referential integrity columns not to conflict
1593 with referential integrity locks
1595 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00073.php
1597 * Add idle_in_transaction_timeout GUC so locks are not held for long
1600 * Improve deadlock detection when a page cleaning lock conflicts
1601 with a shared buffer that is pinned
1603 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00138.php
1604 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1605 http://archives.postgresql.org/pgsql-committers/2008-01/msg00365.php
1607 * Detect deadlocks involving LockBufferForCleanup()
1609 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1613 Startup Time Improvements
1614 =========================
1616 * Experiment with multi-threaded backend for backend creation
1618 This would prevent the overhead associated with process creation. Most
1619 operating systems have trivial process creation time compared to
1620 database startup overhead, but a few operating systems (Win32,
1621 Solaris) might benefit from threading. Also explore the idea of
1622 a single session using multiple threads to execute a statement faster.
1629 * Eliminate need to write full pages to WAL before page modification
1631 Currently, to protect against partial disk page writes, we write
1632 full page images to WAL before they are modified so we can correct any
1633 partial page writes during recovery. These pages can also be
1634 eliminated from point-in-time archive files.
1635 http://archives.postgresql.org/pgsql-hackers/2002-06/msg00655.php
1637 o When off, write CRC to WAL and check file system blocks
1640 If CRC check fails during recovery, remember the page in case
1641 a later CRC for that page properly matches.
1643 o Write full pages during file system write and not when
1644 the page is modified in the buffer cache
1646 This allows most full page writes to happen in the background
1647 writer. It might cause problems for applying WAL on recovery
1648 into a partially-written page, but later the full page will be
1651 * Allow WAL traffic to be streamed to another server for stand-by
1653 * Reduce WAL traffic so only modified values are written rather than
1656 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01589.php
1658 * Allow WAL information to recover corrupted pg_controldata
1660 http://archives.postgresql.org/pgsql-patches/2006-06/msg00025.php
1662 * Find a way to reduce rotational delay when repeatedly writing
1665 Currently fsync of WAL requires the disk platter to perform a full
1666 rotation to fsync again. One idea is to write the WAL to different
1667 offsets that might reduce the rotational delay.
1668 http://archives.postgresql.org/pgsql-hackers/2002-11/msg00483.php
1670 * Allow WAL logging to be turned off for a table, but the table
1671 might be dropped or truncated during crash recovery
1673 Allow tables to bypass WAL writes and just fsync() dirty pages on
1674 commit. This should be implemented using ALTER TABLE, e.g. ALTER
1675 TABLE PERSISTENCE [ DROP | TRUNCATE | DEFAULT ]. Tables using
1676 non-default logging should not use referential integrity with
1677 default-logging tables. A table without dirty buffers during a
1678 crash could perhaps avoid the drop/truncate.
1679 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1681 * Allow WAL logging to be turned off for a table, but the table would
1682 avoid being truncated/dropped
1684 To do this, only a single writer can modify the table, and writes
1685 must happen only on new pages so the new pages can be removed during
1686 crash recovery. Readers can continue accessing the table. Such
1687 tables probably cannot have indexes. One complexity is the handling
1688 of indexes on TOAST tables.
1689 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1691 * Speed WAL recovery by allowing more than one page to be prefetched
1693 This should be done utilizing the same infrastructure used for
1694 prefetching in general to avoid introducing complex error-prone code
1696 http://archives.postgresql.org/pgsql-general/2007-12/msg00683.php
1697 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00497.php
1698 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01279.php
1700 * Improve WAL concurrency by increasing lock granularity
1702 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00556.php
1704 * Be more aggressive about creating WAL files
1706 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01325.php
1707 http://archives.postgresql.org/pgsql-hackers/2004-07/msg01075.php
1708 http://archives.postgresql.org/pgsql-hackers/2005-04/msg00556.php
1710 * Have resource managers report the duration of their status changes
1712 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01468.php
1714 * Move pgfoundry's xlogdump to /contrib and have it rely more closely
1715 on the WAL backend code
1717 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00035.php
1722 Optimizer / Executor
1723 ====================
1725 * Improve selectivity functions for geometric operators
1726 * Precompile SQL functions to avoid overhead
1727 * Create utility to compute accurate random_page_cost value
1728 * Improve ability to display optimizer analysis using OPTIMIZER_DEBUG
1729 * Have EXPLAIN ANALYZE issue NOTICE messages when the estimated and
1730 actual row counts differ by a specified percentage
1731 * Improve how ANALYZE computes in-doubt tuples
1733 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00771.php
1735 * Log statements where the optimizer row estimates were dramatically
1736 different from the number of rows actually found?
1737 * Consider compressed annealing to search for query plans
1739 This might replace GEQO, http://sixdemonbag.org/Djinni.
1741 * Improve merge join performance by allowing mark/restore of
1744 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00096.php
1746 * Consider using a hash for joining to a large IN (VALUES ...) list
1748 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00450.php
1755 * Consider having the background writer update the transaction status
1756 hint bits before writing out the page
1758 Implementing this requires the background writer to have access to system
1759 catalogs and the transaction status log.
1761 * Consider adding buffers the background writer finds reusable to the
1764 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1766 * Automatically tune bgwriter_delay based on activity rather then using a
1769 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1771 * Consider wither increasing BM_MAX_USAGE_COUNT improves performance
1773 http://archives.postgresql.org/pgsql-hackers/2007-06/msg01007.php
1775 * Test to see if calling PreallocXlogFiles() from the background writer
1776 will help with WAL segment creation latency
1778 http://archives.postgresql.org/pgsql-patches/2007-06/msg00340.php
1782 Miscellaneous Performance
1783 =========================
1785 * Do async I/O for faster random read-ahead of data
1787 Async I/O allows multiple I/O requests to be sent to the disk with
1788 results coming back asynchronously.
1790 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00820.php
1791 http://archives.postgresql.org/pgsql-performance/2007-09/msg00255.php
1792 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00027.php
1793 http://archives.postgresql.org/pgsql-patches/2008-01/msg00170.php
1795 * Use mmap() rather than SYSV shared memory or to write WAL files?
1797 This would remove the requirement for SYSV SHM but would introduce
1798 portability issues. Anonymous mmap (or mmap to /dev/zero) is required
1799 to prevent I/O overhead.
1801 * Consider mmap()'ing files into a backend?
1803 Doing I/O to large tables would consume a lot of address space or
1804 require frequent mapping/unmapping. Extending the file also causes
1805 mapping problems that might require mapping only individual pages,
1806 leading to thousands of mappings. Another problem is that there is no
1807 way to _prevent_ I/O to disk from the dirty shared buffers so changes
1808 could hit disk before WAL is written.
1810 * Add a script to ask system configuration questions and tune postgresql.conf
1811 * Consider ways of storing rows more compactly on disk
1813 o Reduce the row header size?
1814 o Consider reducing on-disk varlena length from four bytes to
1815 two because a heap row cannot be more than 64k in length
1817 * Consider increasing NUM_CLOG_BUFFERS
1819 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00030.php
1820 http://archives.postgresql.org/pgsql-performance/2007-08/msg00024.php
1822 * Consider transaction start/end performance improvements
1824 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00948.php
1825 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1827 * Allow user configuration of TOAST thresholds
1829 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00213.php
1830 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00082.php
1832 * Allow configuration of backend priorities via the operating system
1834 Though backend priorities make priority inversion during lock
1835 waits possible, research shows that this is not a huge problem.
1837 http://archives.postgresql.org/pgsql-general/2007-02/msg00493.php
1839 * Experiment with multi-threaded backend better I/O utilization
1841 This would allow a single query to make use of multiple I/O channels
1842 simultaneously. One idea is to create a background reader that can
1843 pre-fetch sequential and index scan pages needed by other backends.
1844 This could be expanded to allow concurrent reads from multiple devices
1845 in a partitioned table.
1847 * Experiment with multi-threaded backend better CPU utilization
1849 This would allow several CPUs to be used for a single query, such as
1850 for sorting or query execution.
1852 * Consider increasing the minimum allowed number of shared buffers
1854 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00157.php
1856 * Expire published xmin for read-only and idle transactions
1858 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00343.php
1860 * Consider if CommandCounterIncrement() can avoid its
1861 AcceptInvalidationMessages() call
1863 http://archives.postgresql.org/pgsql-committers/2007-11/msg00585.php
1865 * Improve performance of shared invalidation queue for multiple CPUs
1867 http://archives.postgresql.org/pgsql-performance/2008-01/msg00023.php
1869 * Consider Cartesian joins when both relations are needed to form an
1870 indexscan qualification for a third relation
1872 http://archives.postgresql.org/pgsql-performance/2007-12/msg00090.php
1874 * Consider not storing a NULL bitmap on disk if all the NULLs are
1877 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00624.php
1878 http://archives.postgresql.org/pgsql-patches/2007-12/msg00109.php
1880 * Sort large UPDATE/DELETEs so it is done in heap order
1882 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01119.php
1884 * SMP scalability improvements
1886 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00439.php
1887 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00206.php
1888 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1890 * Research reducing deTOASTing in more places
1892 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00895.php
1894 * Allow one transaction to see tuples using the snapshot of another
1897 This would assist multiple backends in working together.
1898 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00400.php
1904 * Add use of 'const' for variables in source tree
1905 * Move some things from contrib into main tree
1906 * %Remove warnings created by -Wcast-align
1907 * Move platform-specific ps status display info from ps_status.c to ports
1908 * Add optional CRC checksum to heap and index pages
1909 * Improve documentation to build only interfaces (Marc)
1910 * Remove or relicense modules that are not under the BSD license, if possible
1911 * Acquire lock on a relation before building a relcache entry for it
1912 * Allow cross-compiling by generating the zic database on the target system
1913 * Improve NLS maintenance of libpgport messages linked onto applications
1914 * Clean up casting in contrib/isn
1916 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00245.php
1918 * Improve the /contrib installation experience
1920 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00132.php
1922 * Use UTF8 encoding for NLS messages so all server encodings can
1924 * Update Bonjour to work with newer cross-platform SDK
1926 http://archives.postgresql.org/pgsql-hackers/2006-09/msg02238.php
1927 http://archives.postgresql.org/pgsql-patches/2006-10/msg00048.php
1929 * Consider GnuTLS if OpenSSL license becomes a problem
1931 http://archives.postgresql.org/pgsql-patches/2006-05/msg00040.php
1932 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01213.php
1934 * Consider changing documentation format from SGML to XML
1936 http://archives.postgresql.org/pgsql-docs/2006-12/msg00152.php
1938 * Consider making NAMEDATALEN more configurable in future releases
1939 * Update our code to handle 64-bit timezone files to match the zic
1940 source code, which now uses them
1941 * Have configure choose integer datetimes by default
1943 http://archives.postgresql.org/pgsql-patches/2007-05/msg00046.php
1945 * Support scoped IPv6 addresses
1947 http://archives.postgresql.org/pgsql-bugs/2007-05/msg00111.php
1949 * Consider allowing 64-bit integers and floats to be passed by value on
1952 Also change 32-bit floats (float4) to be passed by value at the same
1955 * Research use of signals and sleep wake ups
1957 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00003.php
1959 * Add automated check for invalid C++ source code constructs
1961 http://archives.postgresql.org/pgsql-patches/2007-07/msg00056.php
1963 * Consider simplifying how memory context resets handle child contexts
1965 http://archives.postgresql.org/pgsql-patches/2007-08/msg00067.php
1967 * Remove use of MAKE_PTR and MAKE_OFFSET macros
1969 http://archives.postgresql.org/pgsql-general/2007-08/msg01510.php
1971 * Convert single quotes to apostrophes in the PDF documentation
1973 http://archives.postgresql.org/pgsql-docs/2007-12/msg00059.php
1975 * Create three versions of libpgport to simplify client code
1977 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00154.php
1979 * Remove old-style routines for manipulating tuples
1981 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00851.php
1983 * Improve detection of shared memory segments being used by others
1984 by checking the SysV shared memory field 'nattch'
1986 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00656.php
1987 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00673.php
1989 * Implement the non-threaded Avahi service discovery protocol
1991 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00939.php
1992 http://archives.postgresql.org/pgsql-patches/2008-02/msg00097.php
1993 http://archives.postgresql.org/pgsql-hackers/2008-03/msg01211.php
1994 http://archives.postgresql.org/pgsql-patches/2008-04/msg00001.php
1996 * Implement a module capability for loading /contrib-style extensions
1998 http://archives.postgresql.org/pgsql-patches/2008-04/msg00164.php
2003 o Remove configure.in check for link failure when cause is found
2004 o Remove readdir() errno patch when runtime/mingwex/dirent.c rev
2006 o -Remove psql newline patch when we find out why mingw outputs an
2008 o Allow psql to use readline once non-US code pages work with
2010 o Fix problem with shared memory on the Win32 Terminal Server
2011 o Diagnose problem where shared memory can sometimes not be
2012 attached by postmaster children
2014 http://archives.postgresql.org/pgsql-general/2007-08/msg01377.php
2016 o Improve signal handling
2018 http://archives.postgresql.org/pgsql-patches/2005-06/msg00027.php
2020 o Convert MSVC build system to remove most batch files
2022 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00961.php
2024 o Prevent SSL from sending network packets to avoid interference
2025 with Win32 signal emulation
2027 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00455.php
2029 o Support pgxs when using MSVC
2031 o Fix MSVC NLS support, like for to_char()
2033 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00485.php
2034 http://archives.postgresql.org/pgsql-patches/2008-02/msg00038.php
2036 o Find a correct rint() substitute on Windows
2038 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00808.php
2042 * Wire Protocol Changes
2044 o Allow dynamic character set handling
2045 o Add decoded type, length, precision
2047 o Update clients to use data types, typmod, schema.table.column names
2048 of result sets using new statement protocol
2054 * Add pre-parsing phase that converts non-ISO syntax to supported
2057 This could allow SQL written for other databases to run without
2060 * Allow plug-in modules to emulate features from other databases
2061 * Add features of Oracle-style packages (Pavel)
2063 A package would be a schema with session-local variables,
2064 public/private functions, and initialization functions. It
2065 is also possible to implement these capabilities
2066 in any schema and not use a separate "packages"
2069 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00384.php
2071 * Consider allowing control of upper/lower case folding of unquoted
2074 http://archives.postgresql.org/pgsql-hackers/2004-04/msg00818.php
2075 http://archives.postgresql.org/pgsql-hackers/2006-10/msg01527.php
2076 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00849.php
2078 * Add autonomous transactions
2080 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00893.php
2084 Features We Do _Not_ Want
2085 =========================
2087 * All backends running as threads in a single process (not wanted)
2089 This eliminates the process protection we get from the current setup.
2090 Thread creation is usually the same overhead as process creation on
2091 modern systems, so it seems unwise to use a pure threaded model.
2093 * Optimizer hints (not wanted)
2095 Optimizer hints are used to work around problems in the optimizer. We
2096 would rather have the problems reported and fixed.
2098 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00506.php
2099 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00517.php
2100 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00663.php
2103 Because we support postfix operators, it isn't possible to make AS
2104 optional and continue to use bison.
2105 http://archives.postgresql.org/pgsql-hackers/2003-04/msg00436.php
2107 http://archives.postgresql.org/pgsql-sql/2006-08/msg00164.php
2109 * Embedded server (not wanted)
2111 While PostgreSQL clients runs fine in limited-resource environments, the
2112 server requires multiple processes and a stable pool of resources to
2113 run reliabily and efficiently. Stripping down the PostgreSQL server
2114 to run in the same process address space as the client application
2115 would add too much complexity and failure cases.