postgresql/doc/TODO

TODO list for PostgreSQL
========================
Last updated:		Wed Feb 28 13:32:12 EST 2001

Current maintainer:	Bruce Momjian (pgman@candle.pha.pa.us)

The most recent version of this document can be viewed at
the PostgreSQL web site, http://www.PostgreSQL.org.

A dash (-) marks changes that will appear in the upcoming 7.1 release.

Names in brackets "[]" indicate more detailed information is available in
the directory pgsql/doc/TODO.detail/ under that name.


RELIABILITY
-----------

PARSER

* SELECT pg_class FROM pg_class generates strange error
* Alter TABLE ADD COLUMN does not honor DEFAULT, add CONSTRAINT
* -Creating index of TIMESTAMP & RELTIME fails, or rename to DATETIME (Thomas)
* -SELECT foo UNION SELECT foo is incorrectly simplified to SELECT foo
* Unique index on base column not honored on inserts from inherited table
  INSERT INTO inherit_table (unique_index_col) VALUES (dup) should fail
  [inheritance] 
* -Be smarter about promoting types when UNION merges different data types
* -redesign INSERT ... SELECT to have two levels of target list
* -have INTERSECT/EXCEPT prevent duplicates unless ALL is specified (Tom)
* -SELECT col::DECIMAL(12,10); fails

VIEWS

* -Views containing aggregates sometimes fail (Jan)
* -Creating view and inheriting the view causes view* to show
  duplicates [inheritance] 
* -Disallow LOCK on view (Mark Hollomon)

MISC

* Plpgsql does not handle quoted mixed-case identifiers
* -Buffer reference counting bugfixes (Tom)
* -Fix libpq bug that causes it to drop backend error message sent
  just before connection closure (ie, any FATAL error message)
* -SELECT ... UNION ... ORDER BY fails when sort expr not in result list
* -SELECT ... UNION ... GROUP BY fails if column types disagree, no type
  promotion occurs
* -Modification  of  pg_class  can  happen while table in use by
  another backend.  Might  lead  to  MVCC  inside  of  syscache
* Permission to DELETE table allows UPDATE also
* SELECT cash_out(1) causes crash

ENHANCEMENTS
------------

URGENT

* -Add OUTER joins, left and right (Tom, Thomas)
* -Allow long tuples by chaining or auto-storing outside db (TOAST) (Jan)
* -Fix memory leak for expressions (Tom) 
* Add replication of distributed databases [replication]
	o automatic fallover
	o load balancing
	o master/slave replication
	o multi-master replication
	o partition data across servers
	o sample implementation in contrib/rserv
	o queries across databases or servers (two-phase commit)
* Point-in-time data recovery using backup and write-ahead log
* Allow row re-use without vacuum (Vadim)

ADMIN

* More access control over who can create tables and use locks (Karel)
* -Test syslog functionality
* Convert remaining fprintf(stderr,...)/perror() to elog()
* Allow elog() to return error codes, module name, file name, line
  number, not just messages
* Allow international error message support and add error codes
* -Unify configuration into one configuration file (Peter E)
* -use setproctitle() if it exists for 'ps' display of status
* -Prevent initdb from running wrong version of postmaster/postgres (Peter E)
* Remove unused files during database vacuum or postmaster startup

TYPES

* Nchar (as distinguished from ordinary varchar),
* Domain capability
* -Add STDDEV/VARIANCE() function for standard deviation computation/variance
* -Allow compression of large fields or a compressed field type
* -Large objects
	o -Fix large object mapping scheme, own typeid or reltype (Peter E)
	o -Not to stuff everything as files in a single directory, hash dirs
	o -Allow large object vacuuming
	o -Tables that start with xinv confused to be large objects
* Add IPv6 capability to INET/CIDR types
* -Fix improper masking of some inet/cidr types 
* Add conversion function from text to inet
* Make a separate SERIAL type?
* Store binary-compatible type information in the system
* -Add support for & operator (Peter E)
* Allow LOCALE on a per-column basis, default to ASCII
* Add SHOW command to see locale
* Allow nulls in arrays
* Allow arrays to be ORDER'ed
* Support construction of array result values in expressions
* Change foreign key constraint for array -> element to mean element
   in array
* Remove Money type, add money formatting for decimal type
* Declare typein/out functions in pg_proc with a special "C string" data type
* Add non-large-object binary field
* Functions returning sets don't really work right[function]
* -Add hash for int8 (Tom)
* SELECT col FROM tab WHERE numeric_col = 10.1 fails
* -Get BIT type working
* Allow better handling of numeric constants, type conversion [typeconv]
* Support multiple simultaneous character sets, per SQL92
* Reject character sequences those are not valid in their charset
* Make functions more multi-byte aware, i.e. trim()
* Make n of CHAR(n)/VARCHAR(n) the number of letters, not bytes
* -Update macaddr manufacturer numbers, or remove the function macaddr_manuf()
* Add btree index support for reltime, tinterval, regproc
* Add rtree index support for line, lseg, path, point
* Make binary/file in/out interface for TOAST columns
* SELECT cash_out(2) crashes because of opaque
* Add SQL standard function bit_length()
* Make oid use unsigned int more reliably, pg_atoi()
* Missing optimizer selectivities for date, r-tree, etc. [optimizer]

VIEWS

* -Allow DISTINCT on views
* -Allow views of aggregate columns
* -Allow views with subselects
* Automatically create rules on views so they are updateable, per SQL92
* -Change elog for complex view ins|upd|del to "cannot {ins|upd|del}
* Add the functionality for "with check option" clause of create view
* Allow NOTIFY in rules involving conditionals
* Evaluate INSERT rules at end of query, rather than beginning

INDEXES

* Allow CREATE INDEX zman_index ON test (date_trunc( 'day', zman ) datetime_ops)
  fails index can't store constant parameters
* -Allow SQL function indexes
* Add FILLFACTOR to index creation
* Re-enable partial indexes
* Allow inherited tables to inherit index
* Prevent pg_attribute from having duplicate oids for indexes (Tom)
* Add UNIQUE capability to non-btree indexes
* Certain indexes will not shrink, i.e. oid indexes with many inserts (Vadim)
* Have UPDATE/DELETE clean out indexes

COMMANDS

* ALTER TABLE ADD COLUMN to inherited table put column in wrong place 
  [inheritance]
* -Add ALTER TABLE ALTER COLUMN feature (Peter E)
* Add ALTER TABLE DROP COLUMN feature [drop]
* -Add ALTER TABLE command to change table ownership (Mark H)
* Add ALTER FUNCTION 
* Add ALTER TABLE ... DROP CONSTRAINT
* -Add ALTER USER command to change user db attributes (Peter E)
* Automatically drop constraints/functions when object is dropped
* CLUSTER
	* cluster all tables at once
	* prent lose of indexes, permissions, inheritance
	* Automatically keep clustering on a table
	* Keep statistics about clustering, perhaps during VACUUM ANALYZE
	  [optimizer]
* Add SIMILAR TO to allow character classes, 'pg_[a-c]%'
* Auto-destroy sequence on DROP of table with SERIAL (Ryan)
* Allow LOCK TABLE tab1, tab2, tab3 so all tables locked in unison [lock]
* Allow INSERT/UPDATE of system-generated oid value for a row
* Allow INSERT INTO my_table VALUES (a, b, c, DEFAULT, x, y, z, ...)
* Allow INSERT INTO tab (col1, ..) VALUES (val1, ..), (val2, ..)
* -Allow multi-level query trees for INSERT INTO ... SELECT
* -Allow ORDER BY...LIMIT in INSERT INTO ... SELECT (Tom)
* -Allow ESCAPE '\' at the end of LIKE for ANSI compliance (Thomas)
* -Rewrite the LIKE handling by rewriting the user string with the 
  supplied ESCAPE
* Allow RULE recompilation
* -Support UNION/INTERSECT/EXCEPT in sub-selects
* -Allow DELETE and UPDATE to use inheritance
* Allow BINARY option to SELECT, like we do with DECLARE
* MOVE 0 should not move to end of cursor
* Overhaul ACL (access control) code
* Add SHOW command to display locks
* Bring INSERT ... VALUES up to full SQL92 spec, disallow missing
  columns, allow DEFAULT
* Allow cursors to be DECLAREd/OPENed/CLOSEed outside transactions
* Allow DELETE WHERE CURRENT OF cursor
* -redesign UNION structures to have separarate target lists

CLIENTS

* Make NULL's come out at the beginning or end depending on the
  ORDER BY direction
* COPY
	* Update reltuples in COPY
	* Allow specification of column names
	* Allow dump/load of CSV format
* fix array handling for ECPG
* -add pg_dump option to dump type names as standard ANSI types
* have pg_dump use LEFT OUTER JOIN in multi-table SELECTs
  or multiple SELECTS to avoid bad system catalog entries
* -make pg_dump dump in oid order, so dependencies are resolved (Philip)
* allow psql \d to show primary and foreign keys
* allow psql \d to show temporary table structure
* add XML interface capability
* -Allow dumping of users/groups separately
* Add a global RESET command for use with connection pooling
* Remove <=6.3 version compatibility in ODBC?
* Add SET REAL_FORMAT and SET DOUBLE_PRECISION_FORMAT using printf args

REFERENTIAL INTEGRITY

* Add MATCH PARTIAL referential integrity
* Check that primary key exists at foreign key definition time
* Prevent column dropping if column is used by foreign key
* Propagate column or table renaming to foreign key constraints
* -Emit a warning at foreign key creation time if no UNIQUE index
  exists on referenced primary key attributes
* Add deferred trigger queue file (Jan)
* Allow oid to act as a foreign key
* Implement dirty reads and use them in RI triggers
* Enforce referential integrity for system tables
* INSERT & UPDATE/DELETE in transaction of primary key fails with 
  deferredTriggerGetPreviousEvent or "change violation" [foreign]
* Add unique indexes to pg_shadow.usename and pg_shadow.usesysid

EXOTIC FEATURES

* Add sql3 recursive unions
* Add the concept of dataspaces/tablespaces [tablespaces]
* Allow queries across multiple databases
* Allow nested transactions (Vadim)
* Allow INSERT/UPDATE ... RETURNING new.col or old.col (Philip)
* SQL*Net listener that makes PostgreSQL appear as an Oracle database 
  to clients
* Incremental backups
* Allow SQL92 schemas
* Handle transaction rollover [transactions]

MISCELLANEOUS

* Increase identifier length(NAMEDATALEN) if small performance hit
* Create a background process for each database that runs while
  database is idle, finding superceeded rows, gathering stats and vacuuming
* Transaction log, so re-do log can be on a separate disk with after-row
  images (Vadim)
* Populate backend status area and write program to dump status data
* Put sort files in their own directory
* Allow autocommit so always in a transaction block
* Show location of syntax error in query [yacc]
* -Redesign the function call interface to handle NULLs better (Tom)
* Overhaul bufmgr/lockmgr/transaction manager
* -Use IPC_EXCL when creating shared memory and semaphores (Tom)
* Encrpyt passwords in pg_shadow table using MD5 (Vince)
* -Use flock() to prevent multiple postmasters on the same port (Tom)
* Allow Java server-side programming [java]
* Add sed-like regular expression search/replace capability

PERFORMANCE
-----------

-FSYNC

* -Allow transaction commits with rollback with no-fsync performance
  (Vadim)
* Delay fsync() when other backends are about to commit too [fsync]

INDEXES

* Use indexes to find min() and max()
* Use index to restrict rows returned by multi-key index when used with
  non-consecutive keys or OR clauses, so fewer heap accesses
* Allow SELECT * FROM tab WHERE int2col = 4 use int2col index, int8,
  float4, numeric/decimal too [optimizer]
* -Include heap CTID in btree index keys, remove equal-key cruft from
  btree (Tom)
* Use indexes with CIDR '<<' (contains) operator
* Allow LIKE indexing optimization for non-ASCII locales
* Be smarter about insertion of already-ordered data into btree index
* Gather more accurate dispersion statistics using indexes
* Add deleted bit to index tuples to reduce heap access
* Prevent index uniqueness checks when UPDATE does not modify column

CACHE

* Cache most recent query plan(s) (Karel) [prepare]
* Shared catalog cache, reduce lseek()'s by caching table size in shared area

VACUUM

* Improve speed with indexes (perhaps recreate index instead) [vacuum]
* Reduce lock time by moving tuples with read lock, then write
  lock and truncate table [vacuum]
* -Redesign ANALYZE in VACUUM so it can be run separately without locks
* Make ANALYZE a separate command
* Allow ANALYZE to ESTIMATE based on certain random precentage of rows
* Add LAZY VACUUM (Vadim)

MISCELLANEOUS

* Allow compression of log and meta data
* Do async I/O to do better read-ahead of data
* Get faster regex() code from Henry Spencer <henry@zoo.utoronto.ca>
  when it is available
* Use mmap() rather than SYSV shared memory(?)
* -Make oid use oidin/oidout not int4in/int4out in pg_type.h (Tom)
* Improve Subplan list handling
* Allow Subplans to use efficient joins(hash, merge) with upper variable
* -use fmgr_info()/fmgr_faddr() instead of fmgr() calls in high-traffic
  places, like GROUP BY, UNIQUE, index processing, etc.
* improve dynamic memory allocation by introducing tuple-context memory
  allocation (Tom)
* -In WHERE tab1.x=3 AND tab1.x=tab2.y, add tab2.y=3
* allow configuration of maximum number of open files
* Remove pg_listener index
* Improve statistics storage in pg_class [performance]
* Add connection pooling [pool]
* Allow persistent backends [persistent]
* Create a transaction processor to aid in persistent connections and
  connection pooling
* Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
  ANALYZE, and CLUSTER
* Force transactions that commit at near the same time use a single fsync()
* Make blind writes go through the file descriptor cache

SOURCE CODE
-----------
* Add use of 'const' for variables in source tree
* -Fix C optimizer problem where fmgr_ptr calls return different types (Tom)
* Does Mariposa source contain any other bug fixes?
* Remove SET KSQO option now that OR processing is improved (Tom)
* -Use macros to define NT open() file parameters, remove NT-specific defines
* -Change CURRENT to OLD internally for rules (Bruce)
* Replace the use of fprint(stderr, ...) with elog() in backend code
* -Allow libedit to be used in place of libreadline
* -BSD/OS does not support locale because there is no LC_MESSAGES (Bruce)
* Remove unused pg_variable, pg_inheritproc, pg_ipl tables
* Experiment with multi-threaded backend [thread]
* Fix problems with libpq non-blocking/async code [async]
* Merge global and template BKI files (Peter E)
* Fix username/password length limits in all areas
* Remove compile-time upper limit on number of backends (MAXBACKENDS) (Tom)
* Make sure all block numbers are unsigned to increase maximum table size
* Merge LockMethodCtl and LockMethodTable into one shared structure (Bruce)
* HOLDER/HOLDERTAB reanme to PROCLOCKLINK/PROCLOCKLINKTAG (Tom)

---------------------------------------------------------------------------


Developers who have claimed items are:
--------------------------------------
	* Billy is Billy G. Allie <Bill.Allie@mug.org>
	* Brook is Brook Milligan <brook@trillium.NMSU.Edu>
	* Bruce is Bruce Momjian<pgman@candle.pha.pa.us>
	* D'Arcy is D'Arcy J.M. Cain <darcy@druid.net>
	* Edmund is Edmund Mergl <E.Mergl@bawue.de>
	* Hiroshi is Hiroshi Inoue <Inoue@tpf.co.jp>
	* Karel is Karel Zak <zakkr@zf.jcu.cz>
	* Jan is Jan Wieck <wieck@sapserv.debis.de>
 	* Marc is Marc Fournier <scrappy@hub.org>
	* Mark is Mark Hollomon <mhh@mindspring.com>
	* Michael is Michael Meskes <meskes@postgresql.org>
	* Oleg is Oleg Bartunov <oleg@sai.msu.su>
	* Peter M is Peter T Mount <peter@retep.org.uk>
	* Peter E is Peter Eisentraut<peter_e@gmx.net>
	* Philip is Philip Warner <pjw@rhyme.com.au>
	* Ross is Ross J. Reedstrom <reedstrm@wallace.ece.rice.edu>
	* Ryan is Ryan Bradetich <rbrad@hpb50023.boi.hp.com>
 	* Stephan is Stephan Szabo <sszabo@megazone23.bigpanda.com>
	* Tatsuo is Tatsuo Ishii <t-ishii@sra.co.jp>
	* Thomas is Thomas Lockhart <lockhart@alumni.caltech.edu>
	* Tom is Tom Lane <tgl@sss.pgh.pa.us>
	* TomH is Tom I Helbekkmo <tih@Hamartun.Priv.NO>
	* Vadim is Vadim B. Mikheev <vadim4o@email.com>