1 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
2 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
5 <title>PostgreSQL TODO List</title>
6 <meta name="generator" content="HTML::TextToHTML v2.25"/>
8 <body bgcolor="#FFFFFF" text="#000000" link="#FF0000" vlink="#A00000" alink="#0000FF">
9 <h1><a name="section_1">PostgreSQL TODO List</a></h1>
10 <p>Current maintainer: Bruce Momjian (<a href="mailto:bruce@momjian.us">bruce@momjian.us</a>)<br/>
11 Last updated: Wed Mar 12 15:41:18 EDT 2008
13 <p>The most recent version of this document can be viewed at<br/>
14 <a href="http://www.postgresql.org/docs/faqs.TODO.html">http://www.postgresql.org/docs/faqs.TODO.html</a>.
16 <p><strong>A hyphen, "-", marks changes that will appear in the upcoming 8.3 release.</strong><br/>
17 <strong>A percent sign, "%", marks items that are easier to implement.</strong>
19 <p>Bracketed items, "[<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?"></a>]", have more detail.
21 <p>This list contains all known PostgreSQL bugs and feature requests. If<br/>
22 you would like to work on an item, please read the Developer's FAQ<br/>
23 first. There is also a developer's wiki at<br/>
24 <a href="http://developer.postgresql.org">http://developer.postgresql.org</a>.
26 <h1><a name="section_2">Administration</a></h1>
29 <li>Allow administrators to safely terminate individual sessions either
30 via an SQL function or SIGTERM
31 <p> Lock table corruption following SIGTERM of an individual backend
32 has been reported in 8.0. A possible cause was fixed in 8.1, but
33 it is unknown whether other problems exist. This item mostly
34 requires additional testing rather than of writing any new code.
36 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-08/msg00174.php">http://archives.postgresql.org/pgsql-hackers/2006-08/msg00174.php</a>
37 <a href="http://archives.postgresql.org/pgsql-hackers/2007-04/msg00218.php">http://archives.postgresql.org/pgsql-hackers/2007-04/msg00218.php</a>
39 </li><li>Check for unreferenced table files created by transactions that were
40 in-progress when the server terminated abruptly
41 <p> <a href="http://archives.postgresql.org/pgsql-patches/2006-06/msg00096.php">http://archives.postgresql.org/pgsql-patches/2006-06/msg00096.php</a>
43 </li><li>Set proper permissions on non-system schemas during db creation
44 <p> Currently all schemas are owned by the super-user because they are copied
45 from the template1 database. However, since all objects are inherited
46 from the template database, it is not clear that setting schemas to the db
49 </li><li>Add function to report the time of the most recent server reload
50 </li><li>Allow statistics collector information to be pulled from the collector
51 process directly, rather than requiring the collector to write a
52 filesystem file twice a second?
53 </li><li>Allow statistics last vacuum/analyze execution times to be displayed
54 without requiring stats_row_level to be enabled
55 </li><li>Allow log_min_messages to be specified on a per-module basis
56 <p> This would allow administrators to see more detailed information from
57 specific sections of the backend, e.g. checkpoints, autovacuum, etc.
58 Another idea is to allow separate configuration files for each module,
59 or allow arbitrary SET commands to be passed to them.
61 </li><li>Simplify ability to create partitioned tables
62 <p> This would allow creation of partitioned tables without requiring
63 creation of triggers or rules for INSERT/UPDATE/DELETE, and constraints
64 for rapid partition selection. Options could include range and hash
67 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg00375.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg00375.php</a>
68 <a href="http://archives.postgresql.org/pgsql-hackers/2007-04/msg00151.php">http://archives.postgresql.org/pgsql-hackers/2007-04/msg00151.php</a>
70 </li><li>Allow auto-selection of partitioned tables for min/max() operations
71 </li><li>Allow more complex user/database default GUC settings
72 <p> Currently ALTER USER and ALTER DATABASE support per-user and
73 per-database defaults. Consider adding per-user-and-database
74 defaults so things like search_path can be defaulted for a
75 specific user connecting to a specific database.
77 </li><li>Allow custom variable classes that can restrict who can set the values
78 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-11/msg00911.php">http://archives.postgresql.org/pgsql-hackers/2006-11/msg00911.php</a>
80 </li><li>Implement the SQL standard mechanism whereby REVOKE ROLE revokes only
81 the privilege granted by the invoking role, and not those granted
83 <p> <a href="http://archives.postgresql.org/pgsql-bugs/2007-05/msg00010.php">http://archives.postgresql.org/pgsql-bugs/2007-05/msg00010.php</a>
85 </li><li>Allow SSL authentication/encryption over unix domain sockets
86 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-12/msg00924.php">http://archives.postgresql.org/pgsql-hackers/2007-12/msg00924.php</a>
88 </li><li>Allow client certificate names to be checked against the client
90 <p> This is already implemented in
91 libpq/fe-secure.c::verify_peer_name_matches_certificate() but the code
94 </li><li>Configuration files
96 <li>Allow pg_hba.conf to specify host names along with IP addresses
97 <p> Host name lookup could occur when the postmaster reads the
98 pg_hba.conf file, or when the backend starts. Another
99 solution would be to reverse lookup the connection IP and
100 check that hostname against the host names in pg_hba.conf.
101 We could also then check that the host name maps to the IP
104 </li><li>%Allow postgresql.conf file values to be changed via an SQL
105 API, perhaps using SET GLOBAL
106 </li><li>Allow the server to be stopped/restarted via an SQL API
107 </li><li>Issue a warning if a change-on-restart-only postgresql.conf value
108 is modified and the server config files are reloaded
109 </li><li>Consider normalizing fractions in postgresql.conf, perhaps
111 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-06/msg00550.php">http://archives.postgresql.org/pgsql-hackers/2007-06/msg00550.php</a>
116 <li>Allow a database in tablespace t1 with tables created in
117 tablespace t2 to be used as a template for a new database created
118 with default tablespace t2
119 <p> Currently all objects in the default database tablespace must
120 have default tablespace specifications. This is because new
121 databases are created by copying directories. If you mix default
122 tablespace tables and tablespace-specified tables in the same
123 directory, creating a new database from such a mixed directory
124 would create a new database with tables that had incorrect
125 explicit tablespaces. To fix this would require modifying
126 pg_class in the newly copied database, which we don't currently
129 </li><li>Allow reporting of which objects are in which tablespaces
130 <p> This item is difficult because a tablespace can contain objects
131 from multiple databases. There is a server-side function that
132 returns the databases which use a specific tablespace, so this
133 requires a tool that will call that function and connect to each
134 database to find the objects in each database for that tablespace.
136 </li><li>Allow WAL replay of CREATE TABLESPACE to work when the directory
137 structure on the recovery computer is different from the original
138 </li><li>Allow per-tablespace quotas
140 </li><li>Point-In-Time Recovery (PITR)
142 <li>Allow a warm standby system to also allow read-only statements
143 [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?pitr">pitr</a>]
144 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg00050.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg00050.php</a>
146 </li><li>%Create dump tool for write-ahead logs for use in determining
147 transaction id for point-in-time recovery
148 <p> This is useful for checking PITR recovery.
150 </li><li>Allow recovery.conf to support the same syntax as
151 postgresql.conf, including quoting
152 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg00497.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg00497.php</a>
156 <h1><a name="section_3">Data Types</a></h1>
159 <li>Change NUMERIC to enforce the maximum precision
160 </li><li>Reduce storage space for small NUMERICs
161 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-02/msg01331.php">http://archives.postgresql.org/pgsql-hackers/2007-02/msg01331.php</a>
162 <a href="http://archives.postgresql.org/pgsql-patches/2007-02/msg00505.php">http://archives.postgresql.org/pgsql-patches/2007-02/msg00505.php</a>
163 <a href="http://archives.postgresql.org/pgsql-hackers/2007-06/msg00715.php">http://archives.postgresql.org/pgsql-hackers/2007-06/msg00715.php</a>
165 </li><li>Fix data types where equality comparison isn't intuitive, e.g. box
166 </li><li>Add support for public SYNONYMs
167 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-03/msg00519.php">http://archives.postgresql.org/pgsql-hackers/2006-03/msg00519.php</a>
169 </li><li>Fix CREATE CAST on DOMAINs
170 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-05/msg00072.php">http://archives.postgresql.org/pgsql-hackers/2006-05/msg00072.php</a>
171 <a href="http://archives.postgresql.org/pgsql-hackers/2006-09/msg01681.php">http://archives.postgresql.org/pgsql-hackers/2006-09/msg01681.php</a>
173 </li><li>Allow domains to be cast
174 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2003-06/msg01206.php">http://archives.postgresql.org/pgsql-hackers/2003-06/msg01206.php</a>
175 <a href="http://archives.postgresql.org/pgsql-hackers/2007-08/msg00289.php">http://archives.postgresql.org/pgsql-hackers/2007-08/msg00289.php</a>
177 </li><li>Add support for SQL-standard GENERATED/IDENTITY columns
178 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-07/msg00543.php">http://archives.postgresql.org/pgsql-hackers/2006-07/msg00543.php</a>
179 <a href="http://archives.postgresql.org/pgsql-hackers/2006-08/msg00038.php">http://archives.postgresql.org/pgsql-hackers/2006-08/msg00038.php</a>
180 <a href="http://archives.postgresql.org/pgsql-hackers/2007-05/msg00344.php">http://archives.postgresql.org/pgsql-hackers/2007-05/msg00344.php</a>
181 <a href="http://archives.postgresql.org/pgsql-patches/2007-05/msg00076.php">http://archives.postgresql.org/pgsql-patches/2007-05/msg00076.php</a>
182 <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/msg00604.php">http://archives.postgresql.org/pgsql-hackers/2008-02/msg00604.php</a>
184 </li><li>Improve XML support
185 <p> <a href="http://developer.postgresql.org/index.php/XML_Support">http://developer.postgresql.org/index.php/XML_Support</a>
187 </li><li>Consider placing all sequences in a single table, or create a system
189 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php">http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php</a>
191 </li><li>Allow the UUID type to accept non-standard formats
192 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/msg01214.php">http://archives.postgresql.org/pgsql-hackers/2008-02/msg01214.php</a>
194 </li><li>Allow text search dictionary to filter out only stop words
195 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-11/msg00081.php">http://archives.postgresql.org/pgsql-patches/2007-11/msg00081.php</a>
197 </li><li>Consider a function-based API for '@@' full text searches
198 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-11/msg00511.php">http://archives.postgresql.org/pgsql-hackers/2007-11/msg00511.php</a>
200 </li><li>Consider a special data type for regular expressions
201 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-08/msg01067.php">http://archives.postgresql.org/pgsql-hackers/2007-08/msg01067.php</a>
203 </li><li>Dates and Times
205 <li>Allow infinite dates and intervals just like infinite timestamps
206 </li><li>Merge hardwired timezone names with the TZ database; allow either
207 kind everywhere a TZ name is currently taken
208 </li><li>Allow TIMESTAMP WITH TIME ZONE to store the original timezone
209 information, either zone name or offset from UTC [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?timezone">timezone</a>]
210 <p> If the TIMESTAMP value is stored with a time zone name, interval
211 computations should adjust based on the time zone rules.
213 </li><li>Fix SELECT '0.01 years'::interval, '0.01 months'::interval
214 </li><li>Add a GUC variable to allow output of interval values in ISO8601
216 </li><li>Have timestamp subtraction not call justify_hours()?
217 <p> <a href="http://archives.postgresql.org/pgsql-sql/2006-10/msg00059.php">http://archives.postgresql.org/pgsql-sql/2006-10/msg00059.php</a>
219 </li><li>Improve timestamptz subtraction to be DST-aware
220 <p> Currently subtracting one date from another that crosses a
221 daylight savings time adjustment can return '1 day 1 hour', but
222 adding that back to the first date returns a time one hour in
223 the future. This is caused by the adjustment of '25 hours' to
224 '1 day 1 hour', and '1 day' is the same time the next day, even
225 if daylight savings adjustments are involved.
227 </li><li>Fix interval display to support values exceeding 2^31 hours
228 </li><li>Add overflow checking to timestamp and interval arithmetic
229 </li><li>Extend timezone code to allow 64-bit values so we can
230 represent years beyond 2038
231 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-09/msg01363.php">http://archives.postgresql.org/pgsql-hackers/2006-09/msg01363.php</a>
233 </li><li>Use LC_TIME for localized weekday/month names, rather than
235 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-11/msg00390.php">http://archives.postgresql.org/pgsql-hackers/2006-11/msg00390.php</a>
237 </li><li>Add ISO INTERVAL handling
240 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-01/msg00250.php">http://archives.postgresql.org/pgsql-hackers/2006-01/msg00250.php</a><br/>
241 <a href="http://archives.postgresql.org/pgsql-bugs/2006-04/msg00248.php">http://archives.postgresql.org/pgsql-bugs/2006-04/msg00248.php</a>
244 <li>Support ISO INTERVAL syntax if units cannot be determined from
245 the string, and are supplied after the string
246 <p> The SQL standard states that the units after the string
247 specify the units of the string, e.g. INTERVAL '2' MINUTE
248 should return '00:02:00'. The current behavior has the units
249 restrict the interval value to the specified unit or unit
250 range, INTERVAL '70' SECOND returns '00:00:10'.
252 <p> For syntax that isn't uniquely ISO or PG syntax, like '1' or
253 '1:30', treat as ISO if there is a range specification clause,
254 and as PG if there no clause is present, e.g. interpret '1:30'
255 MINUTE TO SECOND as '1 minute 30 seconds', and interpret
256 '1:30' as '1 hour, 30 minutes'.
258 <p> This makes common cases like SELECT INTERVAL '1' MONTH
259 SQL-standard results. The SQL standard supports a limited
260 number of unit combinations and doesn't support unit names in
261 the string. The PostgreSQL syntax is more flexible in the
262 range of units supported, e.g. PostgreSQL supports '1 year 1
263 hour', while the SQL standard does not.
265 </li><li>Add support for year-month syntax, INTERVAL '50-6' YEAR
267 </li><li>Interpret INTERVAL '1 year' MONTH as CAST (INTERVAL '1
268 year' AS INTERVAL MONTH), and this should return '12 months'
269 </li><li>Round or truncate values to the requested precision, e.g.
270 INTERVAL '11 months' AS YEAR should return one or zero
271 </li><li>Support precision, CREATE TABLE foo (a INTERVAL MONTH(3))
275 <li>Delay resolution of array expression's data type so assignment
276 coercion can be performed on empty array expressions
277 </li><li>Add support for arrays of domains
278 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-05/msg00114.php">http://archives.postgresql.org/pgsql-patches/2007-05/msg00114.php</a>
280 </li><li>Allow single-byte header storage for array elements
284 <li>Improve vacuum of large objects, like contrib/vacuumlo?
285 </li><li>Add security checking for large objects
286 </li><li>Auto-delete large objects when referencing row is deleted
287 <p> contrib/lo offers this functionality.
289 </li><li>Allow read/write into TOAST values like large objects
290 <p> This requires the TOAST column to be stored EXTERNAL.
292 </li><li>Add API for 64-bit large object access
293 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2005-09/msg00781.php">http://archives.postgresql.org/pgsql-hackers/2005-09/msg00781.php</a>
296 </li><li>MONEY data type
298 <li>Add locale-aware MONEY type, and support multiple currencies
299 <p> <a href="http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php">http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php</a>
300 <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php</a>
302 </li><li>MONEY dumps in a locale-specific format making it difficult to
303 restore to a system with a different locale
304 </li><li>Allow MONEY to be easily cast to/from other numeric data types
308 <h1><a name="section_4">Functions</a></h1>
311 <li>Allow INET subnet tests using non-constants to be indexed
312 </li><li>Allow to_date() and to_timestamp() accept localized month names
313 </li><li>Fix to_date()-related functions to consistently issue errors
314 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-02/msg00915.php">http://archives.postgresql.org/pgsql-hackers/2007-02/msg00915.php</a>
316 </li><li>Add missing parameter handling in to_char()
317 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2005-12/msg00948.php">http://archives.postgresql.org/pgsql-hackers/2005-12/msg00948.php</a>
319 </li><li>Allow substring/replace() to get/set bit values
320 </li><li>Allow to_char() on interval values to accumulate the highest unit
322 <p> Some special format flag would be required to request such
323 accumulation. Such functionality could also be added to EXTRACT.
324 Prevent accumulation that crosses the month/day boundary because of
325 the uneven number of days in a month.
328 <li>to_char(INTERVAL '1 hour 5 minutes', 'MI') => 65
329 </li><li>to_char(INTERVAL '43 hours 20 minutes', 'MI' ) => 2600
330 </li><li>to_char(INTERVAL '43 hours 20 minutes', 'WK:DD:HR:MI') => 0:1:19:20
331 </li><li>to_char(INTERVAL '3 years 5 months','MM') => 41
333 </li><li>Implement inlining of set-returning functions defined in SQL
334 </li><li>Allow SQL-language functions to return results from RETURNING queries
335 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-10/msg00665.php">http://archives.postgresql.org/pgsql-hackers/2006-10/msg00665.php</a>
337 </li><li>Allow SQL-language functions to reference parameters by parameter name
338 <p> Currently SQL-language functions can only refer to dollar parameters,
341 </li><li>Add SPI_gettypmod() to return the typemod for a TupleDesc
342 </li><li>Enforce typmod for function inputs, function results and parameters for
343 spi_prepare'd statements called from PLs
344 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg01403.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg01403.php</a>
346 </li><li>Allow holdable cursors in SPI
347 </li><li>Tighten function permission checks
348 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg00568.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg00568.php</a>
350 </li><li>Fix IS OF so it matches the ISO specification, and add documentation
351 <p> <a href="http://archives.postgresql.org/pgsql-patches/2003-08/msg00060.php">http://archives.postgresql.org/pgsql-patches/2003-08/msg00060.php</a>
352 <a href="http://archives.postgresql.org/pgsql-hackers/2007-02/msg00060.php">http://archives.postgresql.org/pgsql-hackers/2007-02/msg00060.php</a>
354 </li><li>Add missing operators for geometric data types
355 <p> Some geometric types do not have the full suite of geometric operators,
358 </li><li>Implement Boyer-Moore searching in strpos()
359 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-08/msg00012.php">http://archives.postgresql.org/pgsql-patches/2007-08/msg00012.php</a>
361 </li><li>Prevent malicious functions from being executed with the permissions
362 of unsuspecting users
363 <p> Index functions are safe, so VACUUM and ANALYZE are safe too.
364 Triggers, CHECK and DEFAULT expressions, and rules are still vulnerable.
365 <a href="http://archives.postgresql.org/pgsql-hackers/2008-01/msg00268.php">http://archives.postgresql.org/pgsql-hackers/2008-01/msg00268.php</a>
367 </li><li>Reduce memory usage of aggregates in set returning functions
368 <p> <a href="http://archives.postgresql.org/pgsql-performance/2008-01/msg00031.php">http://archives.postgresql.org/pgsql-performance/2008-01/msg00031.php</a>
370 </li><li>Add temporal versions of generate_series()
371 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-04/msg01180.php">http://archives.postgresql.org/pgsql-hackers/2007-04/msg01180.php</a>
373 </li><li>Add array_accum() and array_to_set() functions for arrays
374 <p> The standards specify array_agg() and UNNEST.
375 <a href="http://archives.postgresql.org/pgsql-hackers/2007-08/msg00464.php">http://archives.postgresql.org/pgsql-hackers/2007-08/msg00464.php</a>
378 <h1><a name="section_5">Multi-Language Support</a></h1>
381 <li>Add NCHAR (as distinguished from ordinary varchar),
382 </li><li>Allow locale to be set at database creation
383 <p> Currently locale can only be set during initdb. No global tables have
384 locale-aware columns. However, the database template used during
385 database creation might have locale-aware indexes. The indexes would
386 need to be reindexed to match the new locale.
388 </li><li>Allow encoding on a per-column basis optionally using the ICU library:
389 <p> Right now only one encoding is allowed per database. [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?locale">locale</a>]
391 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2005-03/msg00932.php">http://archives.postgresql.org/pgsql-hackers/2005-03/msg00932.php</a>
392 <a href="http://archives.postgresql.org/pgsql-patches/2005-08/msg00309.php">http://archives.postgresql.org/pgsql-patches/2005-08/msg00309.php</a>
393 <a href="http://archives.postgresql.org/pgsql-patches/2006-03/msg00233.php">http://archives.postgresql.org/pgsql-patches/2006-03/msg00233.php</a>
394 <a href="http://archives.postgresql.org/pgsql-hackers/2006-09/msg00662.php">http://archives.postgresql.org/pgsql-hackers/2006-09/msg00662.php</a>
396 </li><li>Add CREATE COLLATE? [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?locale">locale</a>]
397 </li><li>Support multiple simultaneous character sets, per SQL92
398 </li><li>Improve UTF8 combined character handling?
399 </li><li>Add octet_length_server() and octet_length_client()
400 </li><li>Make octet_length_client() the same as octet_length()?
401 </li><li>Fix problems with wrong runtime encoding conversion for NLS message files
402 </li><li>Add URL to more complete multi-byte regression tests
403 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2005-07/msg00272.php">http://archives.postgresql.org/pgsql-hackers/2005-07/msg00272.php</a>
405 </li><li>Fix ILIKE and regular expressions to handle case insensitivity
406 properly in multibyte encodings
407 <p> <a href="http://archives.postgresql.org/pgsql-bugs/2005-10/msg00001.php">http://archives.postgresql.org/pgsql-bugs/2005-10/msg00001.php</a>
408 <a href="http://archives.postgresql.org/pgsql-patches/2005-11/msg00173.php">http://archives.postgresql.org/pgsql-patches/2005-11/msg00173.php</a>
410 </li><li>Set client encoding based on the client operating system encoding
411 <p> Currently client_encoding is set in postgresql.conf, which
412 defaults to the server encoding.
413 <a href="http://archives.postgresql.org/pgsql-hackers/2006-08/msg01696.php">http://archives.postgresql.org/pgsql-hackers/2006-08/msg01696.php</a>
415 </li><li>Change memory allocation for multi-byte functions so memory is
416 allocated inside conversion functions
417 <p> Currently we preallocate memory based on worst-case usage.
420 <h1><a name="section_6">Views / Rules</a></h1>
423 <li>Automatically create rules on views so they are updateable, per SQL99
424 <p> We can only auto-create rules for simple views. For more complex
425 cases users will still have to write rules manually.
427 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-03/msg00586.php">http://archives.postgresql.org/pgsql-hackers/2006-03/msg00586.php</a>
428 <a href="http://archives.postgresql.org/pgsql-patches/2006-08/msg00255.php">http://archives.postgresql.org/pgsql-patches/2006-08/msg00255.php</a>
430 </li><li>Add the functionality for WITH CHECK OPTION clause of CREATE VIEW
431 </li><li>Allow NOTIFY in rules involving conditionals
432 </li><li>Allow VIEW/RULE recompilation when the underlying tables change
433 <p> Another issue is whether underlying table changes should be reflected
434 in the view, e.g. should SELECT * show additional columns if they
435 are added after the view is created.
437 </li><li>Make it possible to use RETURNING together with conditional DO INSTEAD
438 rules, such as for partitioning setups
439 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-09/msg00577.php">http://archives.postgresql.org/pgsql-hackers/2007-09/msg00577.php</a>
441 </li><li>Add the ability to automatically create materialized views
442 <p> Right now materialized views require the user to create triggers on the
443 main table to keep the summary table current. SQL syntax should be able
444 to manager the triggers and summary table automatically. A more
445 sophisticated implementation would automatically retrieve from the
446 summary table when the main table is referenced, if possible.
449 <h1><a name="section_7">SQL Commands</a></h1>
452 <li>Add CORRESPONDING BY to UNION/INTERSECT/EXCEPT
453 </li><li>Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY
454 </li><li>%Allow SET CONSTRAINTS to be qualified by schema/table name
455 </li><li>%Add a separate TRUNCATE permission
456 <p> Currently only the owner can TRUNCATE a table because triggers are not
457 called, and the table is locked in exclusive mode.
459 </li><li>Allow PREPARE of cursors
460 </li><li>Allow finer control over the caching of prepared query plans
461 <p> Currently queries prepared via the libpq API are planned on first
462 execute using the supplied parameters --- allow SQL PREPARE to do the
463 same. Also, allow control over replanning prepared queries either
464 manually or automatically when statistics for execute parameters
465 differ dramatically from those used during planning.
467 </li><li>Improve logging of prepared transactions recovered during startup
468 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-11/msg00092.php">http://archives.postgresql.org/pgsql-hackers/2006-11/msg00092.php</a>
470 </li><li>Allow LISTEN/NOTIFY to store info in memory rather than tables?
471 <p> Currently LISTEN/NOTIFY information is stored in pg_listener. Storing
472 such information in memory would improve performance.
474 </li><li>Add optional textual message to NOTIFY
475 <p> This would allow an informational message to be added to the notify
476 message, perhaps indicating the row modified or other custom
479 </li><li>Allow multiple identical NOTIFY events to always be communicated to the
480 client, rather than sent as a single notification to the listener
481 <p> <a href="http://archives.postgresql.org/pgsql-general/2008-01/msg00057.php">http://archives.postgresql.org/pgsql-general/2008-01/msg00057.php</a>
483 </li><li>Add a GUC variable to warn about non-standard SQL usage in queries
484 </li><li>Add SQL-standard MERGE command, typically used to merge two tables
485 [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?merge">merge</a>]
486 <p> This is similar to UPDATE, then for unmatched rows, INSERT.
487 Whether concurrent access allows modifications which could cause
488 row loss is implementation independent.
490 </li><li>Add REPLACE or UPSERT command that does UPDATE, or on failure, INSERT
491 [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?merge">merge</a>]
492 <p> To implement this cleanly requires that the table have a unique index
493 so duplicate checking can be easily performed. It is possible to
494 do it without a unique index if we require the user to LOCK the table
497 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2005-11/msg00501.php">http://archives.postgresql.org/pgsql-hackers/2005-11/msg00501.php</a>
498 <a href="http://archives.postgresql.org/pgsql-hackers/2005-11/msg00536.php">http://archives.postgresql.org/pgsql-hackers/2005-11/msg00536.php</a>
500 </li><li>Add NOVICE output level for helpful messages like automatic sequence/index
502 </li><li>Add GUC to issue notice about statements that use unjoined tables
503 </li><li>Allow EXPLAIN to identify tables that were skipped because of
505 </li><li>Allow EXPLAIN output to be more easily processed by scripts, perhaps XML
506 </li><li>Enable standard_conforming_strings
507 </li><li>Make standard_conforming_strings the default in 8.5?
508 <p> When this is done, backslash-quote should be prohibited in non-E''
509 strings because of possible confusion over how such strings treat
510 backslashes. Basically, '' is always safe for a literal single
511 quote, while \' might or might not be based on the backslash
514 </li><li>Simplify dropping roles that have objects in several databases
515 </li><li>Allow COMMENT ON to accept an expression rather than just a string
516 </li><li>Allow the count returned by SELECT, etc to be represented as an int64
517 to allow a higher range of values
518 </li><li>Add SQL99 WITH clause to SELECT
519 </li><li>Add SQL:2003 WITH RECURSIVE (hierarchical) queries to SELECT
520 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg01375.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg01375.php</a>
521 <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/msg00642.php">http://archives.postgresql.org/pgsql-hackers/2008-02/msg00642.php</a>
522 <a href="http://archives.postgresql.org/pgsql-patches/2007-03/msg00139.php">http://archives.postgresql.org/pgsql-patches/2007-03/msg00139.php</a>
524 </li><li>Add DEFAULT .. AS OWNER so permission checks are done as the table
526 <p> This would be useful for SERIAL nextval() calls and CHECK constraints.
528 </li><li>Allow DISTINCT to work in multiple-argument aggregate calls
529 </li><li>Add column to pg_stat_activity that shows the progress of long-running
530 commands like CREATE INDEX and VACUUM
531 </li><li>Implement SQL:2003 window functions
532 </li><li>Improve failure message when DROP DATABASE is used on a database that
533 has prepared transactions
534 </li><li>Allow INSERT/UPDATE ... RETURNING inside a SELECT 'FROM' clause
535 <p> <a href="http://archives.postgresql.org/pgsql-general/2006-09/msg00803.php">http://archives.postgresql.org/pgsql-general/2006-09/msg00803.php</a>
536 <a href="http://archives.postgresql.org/pgsql-hackers/2006-10/msg00693.php">http://archives.postgresql.org/pgsql-hackers/2006-10/msg00693.php</a>
538 </li><li>Increase locking when DROPing objects so dependent objects cannot
539 get dropped while the DROP operation is happening
540 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg00937.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg00937.php</a>
542 </li><li>-<em>Allow AS in "SELECT col AS label" to be optional in certain cases</em>
543 </li><li>Allow INSERT ... DELETE ... RETURNING, namely allow the DELETE ...
544 RETURNING to supply values to the INSERT
545 <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/thrd2.php#00979">http://archives.postgresql.org/pgsql-hackers/2008-02/thrd2.php#00979</a>
546 </li><li>Add comments on system tables/columns using the information in
548 <p> Ideally the information would be pulled from the SGML file
553 <li>Allow CREATE TABLE AS to determine column lengths for complex
554 expressions like SELECT col1 || col2
555 </li><li>Have WITH CONSTRAINTS also create constraint indexes
556 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-04/msg00149.php">http://archives.postgresql.org/pgsql-patches/2007-04/msg00149.php</a>
558 </li><li>Have CONSTRAINT cname NOT NULL record the contraint name
559 <p> Right now pg_attribute.attnotnull records the NOT NULL status
560 of the column, but does not record the contraint name
565 <li>Allow UPDATE tab SET ROW (col, ...) = (SELECT...)
566 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-07/msg01306.php">http://archives.postgresql.org/pgsql-hackers/2006-07/msg01306.php</a>
567 <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg00865.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg00865.php</a>
569 </li><li>Research self-referential UPDATEs that see inconsistent row versions
570 in read-committed mode
571 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-05/msg00507.php">http://archives.postgresql.org/pgsql-hackers/2007-05/msg00507.php</a>
572 <a href="http://archives.postgresql.org/pgsql-hackers/2007-06/msg00016.php">http://archives.postgresql.org/pgsql-hackers/2007-06/msg00016.php</a>
574 </li><li>Allow GLOBAL temporary tables to exist as empty by default in
576 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-07/msg00006.php">http://archives.postgresql.org/pgsql-hackers/2007-07/msg00006.php</a>
581 <li>Have ALTER TABLE RENAME rename SERIAL sequence names
582 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php">http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php</a>
584 </li><li>Have ALTER SEQUENCE RENAME rename the sequence name stored
585 in the sequence table
586 <p> <a href="http://archives.postgresql.org/pgsql-bugs/2007-09/msg00092.php">http://archives.postgresql.org/pgsql-bugs/2007-09/msg00092.php</a>
587 <a href="http://archives.postgresql.org/pgsql-bugs/2007-10/msg00007.php">http://archives.postgresql.org/pgsql-bugs/2007-10/msg00007.php</a>
588 <a href="http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php">http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php</a>
590 </li><li>Add ALTER DOMAIN to modify the underlying data type
591 </li><li>%Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME
592 <p> <a href="http://archives.postgresql.org/pgsql-patches/2006-02/msg00168.php">http://archives.postgresql.org/pgsql-patches/2006-02/msg00168.php</a>
594 </li><li>%Allow ALTER TABLE to change constraint deferrability and actions
595 </li><li>Add missing object types for ALTER ... SET SCHEMA
596 </li><li>Allow ALTER TABLESPACE to move to different directories
597 </li><li>Allow databases to be moved to different tablespaces
598 </li><li>Allow moving system tables to other tablespaces, where possible
599 <p> Currently non-global system tables must be in the default database
600 tablespace. Global system tables can never be moved.
602 </li><li>Prevent parent tables from altering or dropping constraints
603 like CHECK that are inherited by child tables unless CASCADE
605 </li><li>%Prevent child tables from altering or dropping constraints
606 like CHECK that were inherited from the parent table
607 </li><li>Have ALTER INDEX update the name of a constraint using that index
608 </li><li>Add ALTER TABLE RENAME CONSTRAINT, update index name also
609 </li><li>Allow column display reordering by recording a display,
610 storage, and permanent id for every column?
611 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg00782.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg00782.php</a>
616 <li>Automatically maintain clustering on a table
617 <p> This might require some background daemon to maintain clustering
618 during periods of low usage. It might also require tables to be only
619 partially filled for easier reorganization. Another idea would
620 be to create a merged heap/index data file so an index lookup would
621 automatically access the heap data too. A third idea would be to
622 store heap rows in hashed groups, perhaps using a user-supplied
624 <a href="http://archives.postgresql.org/pgsql-performance/2004-08/msg00349.php">http://archives.postgresql.org/pgsql-performance/2004-08/msg00349.php</a>
626 </li><li>%Add default clustering to system tables
627 <p> To do this, determine the ideal cluster index for each system
628 table and set the cluster setting during initdb.
630 </li><li>%Add VERBOSE option to report tables as they are processed,
635 <li>Allow COPY to report error lines and continue
636 <p> This requires the use of a savepoint before each COPY line is
637 processed, with ROLLBACK on COPY failure.
639 </li><li>Allow COPY on a newly-created table to skip WAL logging
640 <p> On crash recovery, the table involved in the COPY would
641 be removed or have its heap and index files truncated. One
642 issue is that no other backend should be able to add to
643 the table at the same time, which is something that is
644 currently allowed. This currently is done if the table is
645 created inside the same transaction block as the COPY because
646 no other backends can see the table.
648 </li><li>Consider using a ring buffer for COPY FROM
649 <p> <a href="http://archives.postgresql.org/pgsql-patches/2008-02/msg00140.php">http://archives.postgresql.org/pgsql-patches/2008-02/msg00140.php</a>
650 <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php">http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php</a>
652 </li><li>Allow COPY FROM to create index entries in bulk
653 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/msg00811.php">http://archives.postgresql.org/pgsql-hackers/2008-02/msg00811.php</a>
655 </li><li>Allow COPY in CSV mode to control whether a quoted zero-length
656 string is treated as NULL
657 <p> Currently this is always treated as a zero-length string,
658 which generates an error when loading into an integer column
659 <a href="http://archives.postgresql.org/pgsql-hackers/2007-07/msg00905.php">http://archives.postgresql.org/pgsql-hackers/2007-07/msg00905.php</a>
662 </li><li>GRANT/REVOKE
664 <li>Allow column-level privileges
665 </li><li>%Allow GRANT/REVOKE permissions to be applied to all schema objects
667 <p> The proposed syntax is:
668 </p><p> GRANT SELECT ON ALL TABLES IN public TO phpuser;
669 GRANT SELECT ON NEW TABLES IN public TO phpuser;
671 </li><li>Allow GRANT/REVOKE permissions to be inherited by objects based on
673 </li><li>Allow SERIAL sequences to inherit permissions from the base table?
677 <li>Prevent DROP TABLE from dropping a row referenced by its own open
682 <li>Allow INSERT/UPDATE of the system-generated oid value for a row
683 </li><li>In rules, allow VALUES() to contain a mixture of 'old' and 'new'
688 <li>Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
692 <h1><a name="section_8">Referential Integrity</a></h1>
695 <li>Add MATCH PARTIAL referential integrity
696 </li><li>Change foreign key constraint for array -> element to mean element
698 </li><li>Fix problem when cascading referential triggers make changes on
699 cascaded tables, seeing the tables in an intermediate state
700 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php">http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php</a>
701 <a href="http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php">http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php</a>
703 </li><li>Allow DEFERRABLE and end-of-statement UNIQUE constraints?
704 <p> This would allow UPDATE tab SET col = col + 1 to work if col has
705 a unique index. Currently, uniqueness checks are done while the
706 command is being executed, rather than at the end of the statement
708 <a href="http://people.planetpostgresql.org/greg/index.php?/archives/2006/06/10.html">http://people.planetpostgresql.org/greg/index.php?/archives/2006/06/10.html</a>
709 <a href="http://archives.postgresql.org/pgsql-hackers/2006-09/msg01458.php">http://archives.postgresql.org/pgsql-hackers/2006-09/msg01458.php</a>
711 </li><li>Optimize referential integrity checks
712 <p> <a href="http://archives.postgresql.org/pgsql-performance/2005-10/msg00458.php">http://archives.postgresql.org/pgsql-performance/2005-10/msg00458.php</a>
713 <a href="http://archives.postgresql.org/pgsql-hackers/2007-04/msg00744.php">http://archives.postgresql.org/pgsql-hackers/2007-04/msg00744.php</a>
716 <h1><a name="section_9">Server-Side Languages</a></h1>
721 <li>Fix RENAME to work on variables other than OLD/NEW
722 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2002-03/msg00591.php">http://archives.postgresql.org/pgsql-hackers/2002-03/msg00591.php</a>
723 <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg01615.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg01615.php</a>
724 <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg01587.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg01587.php</a>
726 </li><li>Allow function parameters to be passed by name,
727 get_employee_salary(12345 AS emp_id, 2001 AS tax_year)
728 </li><li>Allow handling of %TYPE arrays, e.g. tab.col%TYPE[<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?"></a>]
729 </li><li>Allow listing of record column names, and access to
730 record columns via variables, e.g. columns := r.(*),
732 <p> <a href="http://archives.postgresql.org/pgsql-patches/2005-07/msg00458.php">http://archives.postgresql.org/pgsql-patches/2005-07/msg00458.php</a>
733 <a href="http://archives.postgresql.org/pgsql-patches/2006-05/msg00302.php">http://archives.postgresql.org/pgsql-patches/2006-05/msg00302.php</a>
734 <a href="http://archives.postgresql.org/pgsql-patches/2006-06/msg00031.php">http://archives.postgresql.org/pgsql-patches/2006-06/msg00031.php</a>
736 </li><li>Add support for SCROLL cursors
737 </li><li>Add support for WITH HOLD cursors
738 </li><li>Allow row and record variables to be set to NULL constants,
739 and allow NULL tests on such variables
740 <p> Because a row is not scalar, do not allow assignment
741 from NULL-valued scalars.
743 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-10/msg00070.php">http://archives.postgresql.org/pgsql-hackers/2006-10/msg00070.php</a>
745 </li><li>Review handling of MOVE and FETCH
746 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-04/msg00527.php">http://archives.postgresql.org/pgsql-patches/2007-04/msg00527.php</a>
748 </li><li>Improve logic of determining if an identifier is a a
749 variable or column name
750 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-07/msg00436.php">http://archives.postgresql.org/pgsql-hackers/2007-07/msg00436.php</a>
755 <li>Add table function support to pltcl, plpythonu
756 </li><li>Add support for polymorphic arguments and return types to
757 languages other than PL/PgSQL
758 </li><li>Add capability to create and call PROCEDURES
759 </li><li>Add support for OUT and INOUT parameters to languages other
761 </li><li>Add PL/PythonU tracebacks
762 <p> <a href="http://archives.postgresql.org/pgsql-patches/2006-02/msg00288.php">http://archives.postgresql.org/pgsql-patches/2006-02/msg00288.php</a>
764 </li><li>Allow data to be passed in native language formats, rather
766 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-05/msg00289.php">http://archives.postgresql.org/pgsql-hackers/2007-05/msg00289.php</a>
770 <h1><a name="section_10">Clients</a></h1>
773 <li>Have pg_ctl look at PGHOST in case it is a socket directory?
774 </li><li>Allow pg_ctl to work properly with configuration files located outside
776 <p> pg_ctl can not read the pid file because it isn't located in the
777 config directory but in the PGDATA directory. The solution is to
778 allow pg_ctl to read and understand postgresql.conf to find the
779 data_directory value.
781 </li><li>Add a function like pg_get_indexdef() that report more detailed index
783 <p> <a href="http://archives.postgresql.org/pgsql-bugs/2007-12/msg00166.php">http://archives.postgresql.org/pgsql-bugs/2007-12/msg00166.php</a>
785 </li><li>Prefix command-line utilities like createuser with 'pg_'
786 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-06/msg00025.php">http://archives.postgresql.org/pgsql-hackers/2007-06/msg00025.php</a>
790 <li>Have psql show current values for a sequence
791 </li><li>Move psql backslash database information into the backend, use
792 mnemonic commands? [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?psql">psql</a>]
793 <p> This would allow non-psql clients to pull the same information out
794 of the database as psql.
796 </li><li>Fix psql's \d commands more consistent
797 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php">http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php</a>
798 <a href="http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php">http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php</a>
800 </li><li>Consistently display privilege information for all objects in psql
801 </li><li>Add auto-expanded mode so expanded output is used if the row
802 length is wider than the screen width.
803 <p> Consider using auto-expanded mode for backslash commands like \df+.
805 </li><li>Prevent tab completion of SET TRANSACTION from querying the
806 database and therefore preventing the transaction isolation
807 level from being set.
808 <p> Currently SET <tab> causes a database lookup to check all
809 supported session variables. This query causes problems
810 because setting the transaction isolation level must be the
811 first statement of a transaction.
813 </li><li>Add a \set variable to control whether \s displays line numbers
814 <p> Another option is to add \# which lists line numbers, and
815 allows command execution.
817 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg00255.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg00255.php</a>
819 </li><li>Prevent escape string warnings when object names have
821 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2008-01/msg00227.php">http://archives.postgresql.org/pgsql-hackers/2008-01/msg00227.php</a>
823 </li><li>Have \d show foreign keys that reference a table's primary key
824 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-04/msg00424.php">http://archives.postgresql.org/pgsql-hackers/2007-04/msg00424.php</a>
826 </li><li>Have \d show child tables that inherit from the specified parent
827 </li><li>Have \l+ show database size, if permissions allow
828 <p> Ideally it will not generate an error for invalid permissions
831 </li><li>pg_dump / pg_restore
833 <li>%Add dumping of comments on index columns and composite type columns
834 </li><li>%Add full object name to the tag field. eg. for operators we need
835 '=(integer, integer)', instead of just '='.
836 </li><li>Add pg_dumpall custom format dumps?
837 </li><li>Allow selection of individual object(s) of all types, not just
839 </li><li>In a selective dump, allow dumping of an object and all its
841 </li><li>Add options like pg_restore -l and -L to pg_dump
842 </li><li>Stop dumping CASCADE on DROP TYPE commands in clean mode
843 </li><li>Allow pg_dump --clean to drop roles that own objects or have
845 </li><li>Change pg_dump so that a comment on the dumped database is
846 applied to the loaded database, even if the database has a
847 different name. This will require new backend syntax, perhaps
848 COMMENT ON CURRENT DATABASE.
849 </li><li>Remove unnecessary function pointer abstractions in pg_dump source
851 </li><li>Allow pg_dump to utilize multiple CPUs and I/O channels by dumping
852 multiple objects simultaneously
853 <p> The difficulty with this is getting multiple dump processes to
854 produce a single dump output file. It also would require
855 several sessions to share the same snapshot.
856 <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php">http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php</a>
858 </li><li>Allow pg_restore to utilize multiple CPUs and I/O channels by
859 restoring multiple objects simultaneously
860 <p> This might require a pg_restore flag to indicate how many
861 simultaneous operations should be performed. Only pg_dump's
862 -Fc format has the necessary dependency information.
863 <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/msg00963.php">http://archives.postgresql.org/pgsql-hackers/2008-02/msg00963.php</a>
865 </li><li>To better utilize resources, allow pg_restore to check foreign
866 keys simultaneously, where possible
867 </li><li>Allow pg_restore to create all indexes of a table
868 concurrently, via a single heap scan
869 <p> This requires a pg_dump -Fc file because that format contains
870 the required dependency information.
871 <a href="http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php">http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php</a>
873 </li><li>Allow pg_restore to load different parts of the COPY data
875 </li><li>Prevent pg_dump/pg_restore from being affected by
877 <p> Using psql to restore a pg_dump dump is also affected.
883 <p> Document differences between ecpg and the SQL standard and
884 information about the Informix-compatibility module.
886 </li><li>Solve cardinality > 1 for input descriptors / variables?
887 </li><li>Add a semantic check level, e.g. check if a table really exists
888 </li><li>fix handling of DB attributes that are arrays
889 </li><li>Use backend PREPARE/EXECUTE facility for ecpg where possible
890 </li><li>Implement SQLDA
891 </li><li>Fix nested C comments
892 </li><li>%sqlwarn[<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?6">6</a>] should be 'W' if the PRECISION or SCALE value specified
893 </li><li>Make SET CONNECTION thread-aware, non-standard?
894 </li><li>Allow multidimensional arrays
895 </li><li>Add internationalized message strings
896 </li><li>Implement COPY FROM STDIN
900 <li>Add PQescapeIdentifierConn()
901 </li><li>Prevent PQfnumber() from lowercasing unquoted the column name
902 <p> PQfnumber() should never have been doing lowercasing, but
903 historically it has so we need a way to prevent it
905 </li><li>Allow statement results to be automatically batched to the client
906 <p> Currently all statement results are transferred to the libpq
907 client before libpq makes the results available to the
908 application. This feature would allow the application to make
909 use of the first result rows while the rest are transferred, or
910 held on the server waiting for them to be requested by libpq.
911 One complexity is that a statement like SELECT 1/col could error
912 out mid-way through the result set.
914 </li><li>Consider disallowing multiple queries in PQexec() as an
915 additional barrier to SQL injection attacks
916 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg00184.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg00184.php</a>
918 </li><li>Add PQexecf() that allows complex parameter substitution
919 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg01803.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg01803.php</a>
921 </li><li>Add SQLSTATE severity to PGconn return status
922 <p> <a href="http://archives.postgresql.org/pgsql-interfaces/2007-11/msg00015.php">http://archives.postgresql.org/pgsql-interfaces/2007-11/msg00015.php</a>
926 <h1><a name="section_11">Triggers</a></h1>
929 <li>Add deferred trigger queue file
930 <p> Right now all deferred trigger information is stored in backend
931 memory. This could exhaust memory for very large trigger queues.
932 This item involves dumping large queues into files.
934 </li><li>Allow triggers to be disabled in only the current session.
935 <p> This is currently possible by starting a multi-statement transaction,
936 modifying the system tables, performing the desired SQL, restoring the
937 system tables, and committing the transaction. ALTER TABLE ...
938 TRIGGER requires a table lock so it is not ideal for this usage.
940 </li><li>With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY
941 <p> If the dump is known to be valid, allow foreign keys to be added
942 without revalidating the data.
944 </li><li>Allow statement-level triggers to access modified rows
945 </li><li>Support triggers on columns
946 <p> <a href="http://archives.postgresql.org/pgsql-patches/2005-07/msg00107.php">http://archives.postgresql.org/pgsql-patches/2005-07/msg00107.php</a>
948 </li><li>Allow AFTER triggers on system tables
949 <p> System tables are modified in many places in the backend without going
950 through the executor and therefore not causing triggers to fire. To
951 complete this item, the functions that modify system tables will have
954 </li><li>Tighten trigger permission checks
955 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php</a>
957 </li><li>Allow BEFORE INSERT triggers on views
958 <p> <a href="http://archives.postgresql.org/pgsql-general/2007-02/msg01466.php">http://archives.postgresql.org/pgsql-general/2007-02/msg01466.php</a>
960 </li><li>Add ability to trigger on TRUNCATE
961 <p> <a href="http://archives.postgresql.org/pgsql-sql/2008-01/msg00050.php">http://archives.postgresql.org/pgsql-sql/2008-01/msg00050.php</a>
964 <h1><a name="section_12">Indexes</a></h1>
967 <li>Add UNIQUE capability to non-btree indexes
968 </li><li>Prevent index uniqueness checks when UPDATE does not modify the column
969 <p> Uniqueness (index) checks are done when updating a column even if the
970 column is not modified by the UPDATE.
972 </li><li>Allow the creation of on-disk bitmap indexes which can be quickly
973 combined with other bitmap indexes
974 <p> Such indexes could be more compact if there are only a few distinct values.
975 Such indexes can also be compressed. Keeping such indexes updated can be
978 <p> <a href="http://archives.postgresql.org/pgsql-patches/2005-07/msg00512.php">http://archives.postgresql.org/pgsql-patches/2005-07/msg00512.php</a>
979 <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg01107.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg01107.php</a>
981 </li><li>Allow accurate statistics to be collected on indexes with more than
982 one column or expression indexes, perhaps using per-index statistics
983 <p> <a href="http://archives.postgresql.org/pgsql-performance/2006-10/msg00222.php">http://archives.postgresql.org/pgsql-performance/2006-10/msg00222.php</a>
984 <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg01131.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg01131.php</a>
986 </li><li>Consider increasing the number of default statistics target, and
987 reduce statistics target overhead
988 <p> Also consider having a larger statistics target for indexed columns
989 and expression indexes
990 <a href="http://archives.postgresql.org/pgsql-general/2007-05/msg01228.php">http://archives.postgresql.org/pgsql-general/2007-05/msg01228.php</a>
991 <a href="http://archives.postgresql.org/pgsql-general/2007-06/msg00542.php">http://archives.postgresql.org/pgsql-general/2007-06/msg00542.php</a>
993 </li><li>Consider compressing indexes by storing key values duplicated in
994 several rows as a single index entry
995 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg00341.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg00341.php</a>
996 <a href="http://archives.postgresql.org/pgsql-hackers/2007-02/msg01264.php">http://archives.postgresql.org/pgsql-hackers/2007-02/msg01264.php</a>
997 <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg00465.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg00465.php</a>
998 <a href="http://archives.postgresql.org/pgsql-patches/2007-03/msg00163.php">http://archives.postgresql.org/pgsql-patches/2007-03/msg00163.php</a>
1000 </li><li>Add REINDEX CONCURRENTLY, like CREATE INDEX CONCURRENTLY
1001 <p> This is difficult because you must upgrade to an exclusive table lock
1002 to replace the existing index file. CREATE INDEX CONCURRENTLY does not
1003 have this complication. This would allow index compaction without
1005 <a href="http://archives.postgresql.org/pgsql-performance/2007-08/msg00289.php">http://archives.postgresql.org/pgsql-performance/2007-08/msg00289.php</a>
1007 </li><li>Allow multiple indexes to be created concurrently, ideally via a
1008 single heap scan, and have pg_restore use it
1009 </li><li>Inheritance
1011 <li>Allow inherited tables to inherit indexes, UNIQUE constraints,
1012 and primary/foreign keys
1013 </li><li>Honor UNIQUE INDEX on base column in INSERTs/UPDATEs
1014 on inherited table, e.g. INSERT INTO inherit_table
1015 (unique_index_col) VALUES (dup) should fail
1016 <p> The main difficulty with this item is the problem of
1017 creating an index that can span multiple tables.
1019 </li><li>Allow SELECT ... FOR UPDATE on inherited tables
1020 </li><li>Require all CHECK constraints to be inherited
1021 <p> <a href="http://archives.postgresql.org/pgsql-bugs/2007-04/msg00026.php">http://archives.postgresql.org/pgsql-bugs/2007-04/msg00026.php</a>
1026 <li>Add more GIST index support for geometric data types
1027 </li><li>Allow GIST indexes to create certain complex index types, like
1028 digital trees (see Aoki)
1032 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-09/msg00051.php">http://archives.postgresql.org/pgsql-hackers/2007-09/msg00051.php</a>
1035 <li>Pack hash index buckets onto disk pages more efficiently
1036 <p> Currently only one hash bucket can be stored on a page. Ideally
1037 several hash buckets could be stored on a single page and greater
1038 granularity used for the hash algorithm.
1040 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2004-06/msg00168.php">http://archives.postgresql.org/pgsql-hackers/2004-06/msg00168.php</a>
1042 </li><li>Consider sorting hash buckets so entries can be found using a
1043 binary search, rather than a linear scan
1044 </li><li>In hash indexes, consider storing the hash value with or instead
1046 </li><li>Add WAL logging for crash recovery
1047 </li><li>Allow multi-column hash indexes
1048 </li><li>During index creation, pre-sort the tuples to improve build speed
1049 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg01199.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg01199.php</a>
1052 <h1><a name="section_13">Fsync</a></h1>
1055 <li>Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
1056 <p> Ideally this requires a separate test program that can be run
1057 at initdb time or optionally later. Consider O_SYNC when
1060 </li><li>Add program to test if fsync has a delay compared to non-fsync
1061 </li><li>Consider sorting writes during checkpoint
1062 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-06/msg00541.php">http://archives.postgresql.org/pgsql-hackers/2007-06/msg00541.php</a>
1065 <h1><a name="section_14">Cache Usage</a></h1>
1068 <li>Speed up COUNT(*)
1069 <p> We could use a fixed row count and a +/- count to follow MVCC
1070 visibility rules, or a single cached value could be used and
1071 invalidated if anyone modifies the table. Another idea is to
1072 get a count directly from a unique index, but for this to be
1073 faster than a sequential scan it must avoid access to the heap
1074 to obtain tuple visibility information.
1076 </li><li>Provide a way to calculate an "estimated COUNT(*)"
1077 <p> Perhaps by using the optimizer's cardinality estimates or random
1080 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2005-11/msg00943.php">http://archives.postgresql.org/pgsql-hackers/2005-11/msg00943.php</a>
1082 </li><li>Allow data to be pulled directly from indexes
1083 <p> Currently indexes do not have enough tuple visibility information
1084 to allow data to be pulled from the index without also accessing
1085 the heap. One way to allow this is to set a bit on index tuples
1086 to indicate if a tuple is currently visible to all transactions
1087 when the first valid heap lookup happens. This bit would have to
1088 be cleared when a heap tuple is expired.
1090 <p> Another idea is to maintain a bitmap of heap pages where all rows
1091 are visible to all backends, and allow index lookups to reference
1092 that bitmap to avoid heap lookups, perhaps the same bitmap we might
1093 add someday to determine which heap pages need vacuuming. Frequently
1094 accessed bitmaps would have to be stored in shared memory. One 8k
1095 page of bitmaps could track 512MB of heap pages.
1097 <p> A third idea would be for a heap scan to check if all rows are visible
1098 and if so set a per-table flag which can be checked by index scans.
1099 Any change to the table would have to clear the flag. To detect
1100 changes during the heap scan a counter could be set at the start and
1101 checked at the end --- if it is the same, the table has not been
1102 modified --- any table change would increment the counter.
1104 </li><li>Consider automatic caching of statements at various levels:
1106 <li>Parsed query tree
1107 </li><li>Query execute plan
1108 </li><li>Query results
1110 </li><li>Consider increasing internal areas when shared buffers is increased
1111 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2005-10/msg01419.php">http://archives.postgresql.org/pgsql-hackers/2005-10/msg01419.php</a>
1113 </li><li>Consider decreasing the amount of memory used by PrivateRefCount
1114 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-11/msg00797.php">http://archives.postgresql.org/pgsql-hackers/2006-11/msg00797.php</a>
1115 <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg00752.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg00752.php</a>
1117 </li><li>Consider allowing higher priority queries to have referenced buffer
1118 cache pages stay in memory longer
1119 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-11/msg00562.php">http://archives.postgresql.org/pgsql-hackers/2007-11/msg00562.php</a>
1122 <h1><a name="section_15">Vacuum</a></h1>
1125 <li>Improve speed with indexes
1126 <p> For large table adjustments during VACUUM FULL, it is faster to cluster
1127 or reindex rather than update the index. Also, index updates can bloat
1130 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg00024.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg00024.php</a>
1131 <a href="http://archives.postgresql.org/pgsql-performance/2007-05/msg00296.php">http://archives.postgresql.org/pgsql-performance/2007-05/msg00296.php</a>
1132 <a href="http://archives.postgresql.org/pgsql-hackers/2007-08/msg00307.php">http://archives.postgresql.org/pgsql-hackers/2007-08/msg00307.php</a>
1134 </li><li>Auto-fill the free space map by scanning the buffer cache or by
1135 checking pages written by the background writer
1136 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-02/msg01125.php">http://archives.postgresql.org/pgsql-hackers/2006-02/msg01125.php</a>
1137 <a href="http://archives.postgresql.org/pgsql-hackers/2006-03/msg00011.php">http://archives.postgresql.org/pgsql-hackers/2006-03/msg00011.php</a>
1139 </li><li>Create a bitmap of pages that need vacuuming
1140 <p> Instead of sequentially scanning the entire table, have the background
1141 writer or some other process record pages that have expired rows, then
1142 VACUUM can look at just those pages rather than the entire table. In
1143 the event of a system crash, the bitmap would probably be invalidated.
1144 One complexity is that index entries still have to be vacuumed, and
1145 doing this without an index scan (by using the heap values to find the
1146 index entry) might be slow and unreliable, especially for user-defined
1149 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg01188.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg01188.php</a>
1150 <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg00121.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg00121.php</a>
1152 </li><li>Allow FSM to return free space toward the beginning of the heap file,
1153 in hopes that empty pages at the end can be truncated by VACUUM
1154 </li><li>Allow FSM page return free space based on table clustering, to assist
1155 in maintaining clustering?
1156 </li><li>Improve dead row detection during multi-statement transactions usage
1157 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-03/msg00358.php">http://archives.postgresql.org/pgsql-patches/2007-03/msg00358.php</a>
1159 </li><li>Consider a more compact data representation for dead tuples
1160 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-05/msg00143.php">http://archives.postgresql.org/pgsql-patches/2007-05/msg00143.php</a>
1162 </li><li>Auto-vacuum
1164 <li>%Issue log message to suggest VACUUM FULL if a table is nearly
1166 </li><li>Improve control of auto-vacuum
1167 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg00876.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg00876.php</a>
1169 </li><li>Prevent long-lived temporary tables from causing frozen-xid
1170 advancement starvation
1171 <p> The problem is that autovacuum cannot vacuum them to set frozen xids;
1172 only the session that created them can do that.
1173 <a href="http://archives.postgresql.org/pgsql-general/2007-06/msg01645.php">http://archives.postgresql.org/pgsql-general/2007-06/msg01645.php</a>
1175 </li><li>Store per-table autovacuum settings in pg_class.reloptions.
1176 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-02/msg01440.php">http://archives.postgresql.org/pgsql-hackers/2007-02/msg01440.php</a>
1177 <a href="http://archives.postgresql.org/pgsql-hackers/2008-01/msg00724.php">http://archives.postgresql.org/pgsql-hackers/2008-01/msg00724.php</a>
1179 </li><li>Prevent autovacuum from running if an old transaction is still
1180 running from the last vacuum
1181 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-11/msg00899.php">http://archives.postgresql.org/pgsql-hackers/2007-11/msg00899.php</a>
1185 <h1><a name="section_16">Locking</a></h1>
1188 <li>Fix priority ordering of read and write light-weight locks (Neil)
1189 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2004-11/msg00893.php">http://archives.postgresql.org/pgsql-hackers/2004-11/msg00893.php</a>
1190 <a href="http://archives.postgresql.org/pgsql-hackers/2004-11/msg00905.php">http://archives.postgresql.org/pgsql-hackers/2004-11/msg00905.php</a>
1192 </li><li>Fix problem when multiple subtransactions of the same outer transaction
1193 hold different types of locks, and one subtransaction aborts
1194 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-11/msg01011.php">http://archives.postgresql.org/pgsql-hackers/2006-11/msg01011.php</a>
1195 <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg00001.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg00001.php</a>
1196 <a href="http://archives.postgresql.org/pgsql-hackers/2007-02/msg00435.php">http://archives.postgresql.org/pgsql-hackers/2007-02/msg00435.php</a>
1197 <a href="http://archives.postgresql.org/pgsql-hackers/2007-05/msg00773.php">http://archives.postgresql.org/pgsql-hackers/2007-05/msg00773.php</a>
1199 </li><li>Allow UPDATEs on only non-referential integrity columns not to conflict
1200 with referential integrity locks
1201 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-02/msg00073.php">http://archives.postgresql.org/pgsql-hackers/2007-02/msg00073.php</a>
1203 </li><li>Add idle_in_transaction_timeout GUC so locks are not held for long
1205 </li><li>Improve deadlock detection when a page cleaning lock conflicts
1206 with a shared buffer that is pinned
1207 <p> <a href="http://archives.postgresql.org/pgsql-bugs/2008-01/msg00138.php">http://archives.postgresql.org/pgsql-bugs/2008-01/msg00138.php</a>
1208 <a href="http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php">http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php</a>
1209 <a href="http://archives.postgresql.org/pgsql-committers/2008-01/msg00365.php">http://archives.postgresql.org/pgsql-committers/2008-01/msg00365.php</a>
1212 <h1><a name="section_17">Startup Time Improvements</a></h1>
1215 <li>Experiment with multi-threaded backend for backend creation [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?thread">thread</a>]
1216 <p> This would prevent the overhead associated with process creation. Most
1217 operating systems have trivial process creation time compared to
1218 database startup overhead, but a few operating systems (Win32,
1219 Solaris) might benefit from threading. Also explore the idea of
1220 a single session using multiple threads to execute a statement faster.
1223 <h1><a name="section_18">Write-Ahead Log</a></h1>
1226 <li>Eliminate need to write full pages to WAL before page modification [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?wal">wal</a>]
1227 <p> Currently, to protect against partial disk page writes, we write
1228 full page images to WAL before they are modified so we can correct any
1229 partial page writes during recovery. These pages can also be
1230 eliminated from point-in-time archive files.
1233 <li>When off, write CRC to WAL and check file system blocks
1235 <p> If CRC check fails during recovery, remember the page in case
1236 a later CRC for that page properly matches.
1238 </li><li>Write full pages during file system write and not when
1239 the page is modified in the buffer cache
1240 <p> This allows most full page writes to happen in the background
1241 writer. It might cause problems for applying WAL on recovery
1242 into a partially-written page, but later the full page will be
1246 </li><li>Allow WAL traffic to be streamed to another server for stand-by
1248 </li><li>Reduce WAL traffic so only modified values are written rather than
1250 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-03/msg01589.php">http://archives.postgresql.org/pgsql-hackers/2007-03/msg01589.php</a>
1252 </li><li>Allow WAL information to recover corrupted pg_controldata
1253 <p> <a href="http://archives.postgresql.org/pgsql-patches/2006-06/msg00025.php">http://archives.postgresql.org/pgsql-patches/2006-06/msg00025.php</a>
1255 </li><li>Find a way to reduce rotational delay when repeatedly writing
1257 <p> Currently fsync of WAL requires the disk platter to perform a full
1258 rotation to fsync again. One idea is to write the WAL to different
1259 offsets that might reduce the rotational delay.
1261 </li><li>Allow WAL logging to be turned off for a table, but the table
1262 might be dropped or truncated during crash recovery [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?walcontrol">walcontrol</a>]
1263 <p> Allow tables to bypass WAL writes and just fsync() dirty pages on
1264 commit. This should be implemented using ALTER TABLE, e.g. ALTER
1265 TABLE PERSISTENCE [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo? DROP | TRUNCATE | DEFAULT "> DROP | TRUNCATE | DEFAULT </a>]. Tables using
1266 non-default logging should not use referential integrity with
1267 default-logging tables. A table without dirty buffers during a
1268 crash could perhaps avoid the drop/truncate.
1270 </li><li>Allow WAL logging to be turned off for a table, but the table would
1271 avoid being truncated/dropped [<a href="http://momjian.postgresql.org/cgi-bin/pgtodo?walcontrol">walcontrol</a>]
1272 <p> To do this, only a single writer can modify the table, and writes
1273 must happen only on new pages so the new pages can be removed during
1274 crash recovery. Readers can continue accessing the table. Such
1275 tables probably cannot have indexes. One complexity is the handling
1276 of indexes on TOAST tables.
1278 </li><li>Speed WAL recovery by allowing more than one page to be prefetched
1279 <p> This involves having a separate process that can be told which pages
1280 the recovery process will need in the near future.
1281 <a href="http://archives.postgresql.org/pgsql-hackers/2008-02/msg01279.php">http://archives.postgresql.org/pgsql-hackers/2008-02/msg01279.php</a>
1284 <h1><a name="section_19">Optimizer / Executor</a></h1>
1287 <li>Improve selectivity functions for geometric operators
1288 </li><li>Precompile SQL functions to avoid overhead
1289 </li><li>Create utility to compute accurate random_page_cost value
1290 </li><li>Improve ability to display optimizer analysis using OPTIMIZER_DEBUG
1291 </li><li>Have EXPLAIN ANALYZE issue NOTICE messages when the estimated and
1292 actual row counts differ by a specified percentage
1293 </li><li>Consider using hash buckets to do DISTINCT, rather than sorting
1294 <p> This would be beneficial when there are few distinct values. This is
1295 already used by GROUP BY.
1297 </li><li>Log statements where the optimizer row estimates were dramatically
1298 different from the number of rows actually found?
1299 </li><li>Consider compressed annealing to search for query plans
1300 <p> This might replace GEQO, <a href="http://sixdemonbag.org/Djinni">http://sixdemonbag.org/Djinni</a>.
1302 </li><li>Improve merge join performance by allowing mark/restore of
1304 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-01/msg00096.php">http://archives.postgresql.org/pgsql-hackers/2007-01/msg00096.php</a>
1306 </li><li>Consider using a hash for joining to a large IN (VALUES ...) list
1307 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-05/msg00450.php">http://archives.postgresql.org/pgsql-hackers/2007-05/msg00450.php</a>
1310 <h1><a name="section_20">Background Writer</a></h1>
1313 <li>Consider having the background writer update the transaction status
1314 hint bits before writing out the page
1315 <p> Implementing this requires the background writer to have access to system
1316 catalogs and the transaction status log.
1318 </li><li>Consider adding buffers the background writer finds reusable to the
1320 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php">http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php</a>
1322 </li><li>Automatically tune bgwriter_delay based on activity rather then using a
1324 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php">http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php</a>
1326 </li><li>Consider wither increasing BM_MAX_USAGE_COUNT improves performance
1327 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-06/msg01007.php">http://archives.postgresql.org/pgsql-hackers/2007-06/msg01007.php</a>
1329 </li><li>Test to see if calling PreallocXlogFiles() from the background writer
1330 will help with WAL segment creation latency
1331 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-06/msg00340.php">http://archives.postgresql.org/pgsql-patches/2007-06/msg00340.php</a>
1334 <h1><a name="section_21">Miscellaneous Performance</a></h1>
1337 <li>Do async I/O for faster random read-ahead of data
1338 <p> Async I/O allows multiple I/O requests to be sent to the disk with
1339 results coming back asynchronously.
1341 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-10/msg00820.php">http://archives.postgresql.org/pgsql-hackers/2006-10/msg00820.php</a>
1342 <a href="http://archives.postgresql.org/pgsql-hackers/2007-12/msg00027.php">http://archives.postgresql.org/pgsql-hackers/2007-12/msg00027.php</a>
1344 </li><li>Use mmap() rather than SYSV shared memory or to write WAL files?
1345 <p> This would remove the requirement for SYSV SHM but would introduce
1346 portability issues. Anonymous mmap (or mmap to /dev/zero) is required
1347 to prevent I/O overhead.
1349 </li><li>Consider mmap()'ing files into a backend?
1350 <p> Doing I/O to large tables would consume a lot of address space or
1351 require frequent mapping/unmapping. Extending the file also causes
1352 mapping problems that might require mapping only individual pages,
1353 leading to thousands of mappings. Another problem is that there is no
1354 way to <u>prevent</u> I/O to disk from the dirty shared buffers so changes
1355 could hit disk before WAL is written.
1357 </li><li>Add a script to ask system configuration questions and tune postgresql.conf
1358 </li><li>Consider ways of storing rows more compactly on disk
1360 <li>Reduce the row header size?
1361 </li><li>Consider reducing on-disk varlena length from four bytes to
1362 two because a heap row cannot be more than 64k in length
1364 </li><li>Consider increasing NUM_CLOG_BUFFERS
1365 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-08/msg00030.php">http://archives.postgresql.org/pgsql-hackers/2007-08/msg00030.php</a>
1366 <a href="http://archives.postgresql.org/pgsql-performance/2007-08/msg00024.php">http://archives.postgresql.org/pgsql-performance/2007-08/msg00024.php</a>
1368 </li><li>Allow user configuration of TOAST thresholds
1369 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-02/msg00213.php">http://archives.postgresql.org/pgsql-hackers/2007-02/msg00213.php</a>
1370 <a href="http://archives.postgresql.org/pgsql-hackers/2007-08/msg00082.php">http://archives.postgresql.org/pgsql-hackers/2007-08/msg00082.php</a>
1372 </li><li>Allow configuration of backend priorities via the operating system
1373 <p> Though backend priorities make priority inversion during lock
1374 waits possible, research shows that this is not a huge problem.
1376 <p> <a href="http://archives.postgresql.org/pgsql-general/2007-02/msg00493.php">http://archives.postgresql.org/pgsql-general/2007-02/msg00493.php</a>
1378 </li><li>Experiment with multi-threaded backend better resource utilization
1379 <p> This would allow a single query to make use of multiple CPU's or
1380 multiple I/O channels simultaneously. One idea is to create a
1381 background reader that can pre-fetch sequential and index scan
1382 pages needed by other backends. This could be expanded to allow
1383 concurrent reads from multiple devices in a partitioned table.
1385 </li><li>Consider increasing the minimum allowed number of shared buffers
1386 <p> <a href="http://archives.postgresql.org/pgsql-bugs/2008-02/msg00157.php">http://archives.postgresql.org/pgsql-bugs/2008-02/msg00157.php</a>
1389 <h1><a name="section_22">Source Code</a></h1>
1392 <li>Add use of 'const' for variables in source tree
1393 </li><li>Move some things from contrib into main tree
1394 </li><li>%Remove warnings created by -Wcast-align
1395 </li><li>Move platform-specific ps status display info from ps_status.c to ports
1396 </li><li>Add optional CRC checksum to heap and index pages
1397 </li><li>Improve documentation to build only interfaces (Marc)
1398 </li><li>Remove or relicense modules that are not under the BSD license, if possible
1399 </li><li>Acquire lock on a relation before building a relcache entry for it
1400 </li><li>Allow cross-compiling by generating the zic database on the target system
1401 </li><li>Improve NLS maintenance of libpgport messages linked onto applications
1402 </li><li>Clean up casting in contrib/isn
1403 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-11/msg00245.php">http://archives.postgresql.org/pgsql-hackers/2006-11/msg00245.php</a>
1405 </li><li>Use UTF8 encoding for NLS messages so all server encodings can
1407 </li><li>Update Bonjour to work with newer cross-platform SDK
1408 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-09/msg02238.php">http://archives.postgresql.org/pgsql-hackers/2006-09/msg02238.php</a>
1409 <a href="http://archives.postgresql.org/pgsql-patches/2006-10/msg00048.php">http://archives.postgresql.org/pgsql-patches/2006-10/msg00048.php</a>
1411 </li><li>Consider detoasting keys before sorting
1412 </li><li>Consider GnuTLS if OpenSSL license becomes a problem
1413 <p> <a href="http://archives.postgresql.org/pgsql-patches/2006-05/msg00040.php">http://archives.postgresql.org/pgsql-patches/2006-05/msg00040.php</a>
1414 <a href="http://archives.postgresql.org/pgsql-hackers/2006-12/msg01213.php">http://archives.postgresql.org/pgsql-hackers/2006-12/msg01213.php</a>
1416 </li><li>Consider changing documentation format from SGML to XML
1417 <p> <a href="http://archives.postgresql.org/pgsql-docs/2006-12/msg00152.php">http://archives.postgresql.org/pgsql-docs/2006-12/msg00152.php</a>
1419 </li><li>Consider making NAMEDATALEN more configurable in future releases
1420 </li><li>Update our code to handle 64-bit timezone files to match the zic
1421 source code, which now uses them
1422 </li><li>Have configure choose integer datetimes by default
1423 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-05/msg00046.php">http://archives.postgresql.org/pgsql-patches/2007-05/msg00046.php</a>
1425 </li><li>Support scoped IPv6 addresses
1426 <p> <a href="http://archives.postgresql.org/pgsql-bugs/2007-05/msg00111.php">http://archives.postgresql.org/pgsql-bugs/2007-05/msg00111.php</a>
1428 </li><li>Consider allowing 64-bit integers and floats to be passed by value on
1430 <p> Also change 32-bit floats (float4) to be passed by value at the same
1433 </li><li>Research use of signals and sleep wake ups
1434 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-07/msg00003.php">http://archives.postgresql.org/pgsql-hackers/2007-07/msg00003.php</a>
1436 </li><li>Add automated check for invalid C++ source code constructs
1437 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-07/msg00056.php">http://archives.postgresql.org/pgsql-patches/2007-07/msg00056.php</a>
1439 </li><li>Consider simplifying how memory context resets handle child contexts
1440 <p> <a href="http://archives.postgresql.org/pgsql-patches/2007-08/msg00067.php">http://archives.postgresql.org/pgsql-patches/2007-08/msg00067.php</a>
1442 </li><li>Remove use of MAKE_PTR and MAKE_OFFSET macros
1443 <p> <a href="http://archives.postgresql.org/pgsql-general/2007-08/msg01510.php">http://archives.postgresql.org/pgsql-general/2007-08/msg01510.php</a>
1447 <li>Remove configure.in check for link failure when cause is found
1448 </li><li>Remove readdir() errno patch when runtime/mingwex/dirent.c rev
1450 </li><li>Remove psql newline patch when we find out why mingw outputs an
1452 </li><li>Allow psql to use readline once non-US code pages work with
1454 </li><li>Fix problem with shared memory on the Win32 Terminal Server
1455 </li><li>Diagnose problem where shared memory can sometimes not be
1456 attached by postmaster children
1457 <p> <a href="http://archives.postgresql.org/pgsql-general/2007-08/msg01377.php">http://archives.postgresql.org/pgsql-general/2007-08/msg01377.php</a>
1459 </li><li>Improve signal handling
1460 <p> <a href="http://archives.postgresql.org/pgsql-patches/2005-06/msg00027.php">http://archives.postgresql.org/pgsql-patches/2005-06/msg00027.php</a>
1462 </li><li>Convert MSVC build system to remove most batch files
1463 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2007-08/msg00961.php">http://archives.postgresql.org/pgsql-hackers/2007-08/msg00961.php</a>
1466 </li><li>Wire Protocol Changes
1468 <li>Allow dynamic character set handling
1469 </li><li>Add decoded type, length, precision
1470 </li><li>Use compression?
1471 </li><li>Update clients to use data types, typmod, schema.table.column names
1472 of result sets using new statement protocol
1475 <h1><a name="section_23">Exotic Features</a></h1>
1478 <li>Add pre-parsing phase that converts non-ISO syntax to supported
1480 <p> This could allow SQL written for other databases to run without
1483 </li><li>Allow plug-in modules to emulate features from other databases
1484 </li><li>Add features of Oracle-style packages (Pavel)
1485 <p> A package would be a schema with session-local variables,
1486 public/private functions, and initialization functions. It
1487 is also possible to implement these capabilities
1488 in any schema and not use a separate "packages"
1491 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-08/msg00384.php">http://archives.postgresql.org/pgsql-hackers/2006-08/msg00384.php</a>
1493 </li><li>Consider allowing control of upper/lower case folding of unquoted
1495 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2004-04/msg00818.php">http://archives.postgresql.org/pgsql-hackers/2004-04/msg00818.php</a>
1496 <a href="http://archives.postgresql.org/pgsql-hackers/2006-10/msg01527.php">http://archives.postgresql.org/pgsql-hackers/2006-10/msg01527.php</a>
1498 </li><li>Add autonomous transactions
1499 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2008-01/msg00893.php">http://archives.postgresql.org/pgsql-hackers/2008-01/msg00893.php</a>
1502 <h1><a name="section_24">Features We Do <u>Not</u> Want</a></h1>
1505 <li>All backends running as threads in a single process (not wanted)
1506 <p> This eliminates the process protection we get from the current setup.
1507 Thread creation is usually the same overhead as process creation on
1508 modern systems, so it seems unwise to use a pure threaded model.
1510 </li><li>Optimizer hints (not wanted)
1511 <p> Optimizer hints are used to work around problems in the optimizer. We
1512 would rather have the problems reported and fixed.
1514 <p> <a href="http://archives.postgresql.org/pgsql-hackers/2006-08/msg00506.php">http://archives.postgresql.org/pgsql-hackers/2006-08/msg00506.php</a>
1515 <a href="http://archives.postgresql.org/pgsql-hackers/2006-10/msg00517.php">http://archives.postgresql.org/pgsql-hackers/2006-10/msg00517.php</a>
1516 <a href="http://archives.postgresql.org/pgsql-hackers/2006-10/msg00663.php">http://archives.postgresql.org/pgsql-hackers/2006-10/msg00663.php</a>
1518 <p> Because we support postfix operators, it isn't possible to make AS
1519 optional and continue to use bison.
1520 <a href="http://archives.postgresql.org/pgsql-hackers/2003-04/msg00436.php">http://archives.postgresql.org/pgsql-hackers/2003-04/msg00436.php</a>
1522 <p> <a href="http://archives.postgresql.org/pgsql-sql/2006-08/msg00164.php">http://archives.postgresql.org/pgsql-sql/2006-08/msg00164.php</a>
1524 </li><li>Embedded server (not wanted)
1525 <p> While PostgreSQL clients runs fine in limited-resource environments, the
1526 server requires multiple processes and a stable pool of resources to
1527 run reliabily and efficiently. Stripping down the PostgreSQL server
1528 to run in the same process address space as the client application
1529 would add too much complexity and failure cases.</p>