<HTML>
<HEAD>
<TITLE>PostgreSQL TODO list</TITLE>
<BODY BGCOLOR="#FFFFFF" TEXT="#000000" LINK="#FF0000" VLINK="#A00000"\
ALINK="#0000FF">
<META NAME="generator" CONTENT="txt2html v1.25">
</HEAD>
<BODY>
<H1><A NAME="section-1">TODO list for PostgreSQL</A></H1>
Last updated: Tue Sep 28 00:34:21 EDT 1999
<P>
Current maintainer: Bruce Momjian (<A HREF="mailto:maillist@candle.pha.pa.us">maillist@candle.pha.pa.us</A>)
<P>
The most recent version of this document can be viewed at<BR>
the PostgreSQL web site, <A HREF="http://www.PostgreSQL.org">http://www.PostgreSQL.org</A>.
<P>
A dash(-) marks changes that will appear in the next release.
<P>
Names in brackets "[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/"></A>]" indicate more detailed information is available in<BR>
the directory pgsql/doc/TODO.detail/ under that name.
<H2><A NAME="section-1.1">RELIABILITY</A></H2>
<P>
<STRONG>RESOURCES</STRONG>
<UL>
<LI> Elog() does not free all its memory(Jan)
<LI> spinlock stuck problem when elog(FATAL) and elog(ERROR) inside bufmgr
<LI> Recover or force failure when disk space is exhausted
</UL>
<P>
<STRONG>PARSER</STRONG>
<UL>
<LI> Disallow inherited columns with the same name as new columns
<LI> INSERT INTO ... SELECT with AS columns matching result columns problem
<LI> SELECT pg<U>class FROM pg</U>class generates strange error
<LI> Alter TABLE ADD COLUMN does not honor DEFAULT, add CONSTRAINT
<LI> Do not allow bpchar column creation without length
<LI> -Select a[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/1">1</A>] FROM test fails, it needs test.a[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/1">1</A>]
<LI> -Array index references without table name cause problems [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/array">array</A>]
<LI> Update table SET table.value = 3 fails(SQL standard says this is OK)
<LI> Creating index of TIMESTAMP & RELTIME fails, or rename to DATETIME(Thomas)
<LI> SELECT foo UNION SELECT foo is incorrectly simplified to SELECT foo
<LI> -INSERT ... SELECT ... GROUP BY groups by target columns not source columns
<LI> -CREATE TABLE test (a char(5) DEFAULT text '', b int4) fails on INSERT
<LI> UNION with LIMIT fails
<LI> Unique index on base column not honored on inserts from inherited table
INSERT INTO inherit_table (unique<U>index</U>col) VALUES (dup) should fail
[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/inherit">inherit</A>]
<LI> CREATE TABLE x AS SELECT 1 UNION SELECT 2 fails
<LI> CREATE TABLE test(col char(2) DEFAULT user) fails in length restriction
<LI> mismatched types in CREATE TABLE ... DEFAULT causes problems [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/default">default</A>]
<LI> SELECT ... UNION ... ORDER BY fails when sort expr not in result list
<LI> Be smarter about promoting types when UNION merges different data types
<LI> SELECT ... UNION ... GROUP BY fails if column types disagree
<LI> redesign INSERT ... SELECT to have two levels of target list
<LI> -select * from pg_class where oid in (0,-1)
<LI> have INTERSECT/EXCEPT prevent duplicates unless ALL is specified
<LI> prevent primary key of nine columns [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/primary">primary</A>]
<LI> SELECT COUNT('asdf') FROM pg_class WHERE oid=12 crashes
<LI> SELECT DISTINCT ON col1 col1 col2 FROM tab1 is broken [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/distinct">distinct</A>]
<LI> -When using aggregates + GROUP BY, no rows in should yield no rows out
</UL>
<P>
<STRONG>VIEWS</STRONG>
<UL>
<LI> Views containing aggregates sometimes fail(Jan)
<LI> Views with spaces in view name fail when referenced
</UL>
<P>
<STRONG>MISC</STRONG>
<UL>
<LI> User who can create databases can modify pg_database table
<LI> Plpgsql does not handle quoted mixed-case identifiers
<LI> Fix btree to give a useful elog when key > 1/2 (page - overhead)
<LI> pg_dump should preserve primary key information
<LI> plpgsql regression tests fail on BSD/OS
</UL>
<H2><A NAME="section-1.2">ENHANCEMENTS</A></H2>
<P>
<STRONG>URGENT</STRONG>
<UL>
<LI> Add referential integrity(Jan?)[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/primary">primary</A>]
<LI> Add OUTER joins, left and right[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/outer">outer</A>](Thomas, Bruce)
<LI> Allow long tuples by chaining or auto-storing outside db (chaining,large objs)
<LI> Eliminate limits on query length
<LI> Fix memory leak for expressions?[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/memory">memory</A>](Tom?)
<LI> -Fix memory leak for aggregates?
</UL>
<P>
<STRONG>ADMIN</STRONG>
<UL>
<LI> Better interface for adding to pg_group
<LI> More access control over who can create tables and access the database
<LI> Test syslog functionality
<LI> Allow elog() to return error codes, not just messages
<LI> Allow international error message support and add error codes
<LI> Generate postmaster pid file and remove flock/fcntl lock code [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/flock">flock</A>]
<LI> Add ability to specifiy location of lock/socket files [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/flock">flock</A>]
</UL>
<P>
<STRONG>TYPES</STRONG>
<UL>
<LI> Add BIT, BIT VARYING
<LI> Nchar (as distinguished from ordinary varchar),
<LI> Domain capability
<LI> Add STDDEV/VARIANCE() function for standard deviation computation/variance
<LI> Allow compression of large fields or a compressed field type
<LI> Large objects
<UL>
<LI> Fix large object mapping scheme, own typeid or reltype(Peter)
<LI> Allow large text type to use large objects(Peter)
<LI> Not to stuff everything as files in a single directory, hash dirs
<LI> Allow large object vacuuming
<LI> Tables that start with xinv confused to be large objects
</UL>
<LI> Allow pg_descriptions when creating types, tables, columns, and functions
<LI> Add IPv6 capability to INET/CIDR types
<LI> Make a separate SERIAL type?
<LI> Store binary-compatible type information in the system
<LI> Allow user to define char1 column
<LI> Add support for & operator
<LI> Allow LOCALE on a per-column basis, default to ASCII
<LI> Allow array on int8[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/"></A>]
<LI> Allow nulls in arrays
<LI> Allow arrays to be ORDER'ed
<LI> Remove Money type, add money formatting for decimal type
<LI> Declare typein/out functions in pg_proc with a special "C string" data type
<LI> Add non-large-object binary field
<LI> Add index on NUMERIC/DECIMAL type
<LI> Make Absolutetime/Relativetime int4 because time_t can be int8 on some ports
<LI> Functions returning sets don't really work right[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/function">function</A>]
</UL>
<P>
<STRONG>VIEWS</STRONG>
<UL>
<LI> Allow DISTINCT on views
<LI> Allow views of aggregate columns
<LI> Allow views with subselects
</UL>
<P>
<STRONG>INDEXES</STRONG>
<UL>
<LI> Allow CREATE INDEX zman_index ON test (date_trunc( 'day', zman ) datetime_ops)
fails index can't store constant parameters
<LI> Allow creation of functional indexes to use default types
<LI> Permissions on indexes - prevent them?
<LI> Allow SQL function indexes
<LI> Add FILLFACTOR to index creation
<LI> Allow indexing of LIKE with localle character sets
<LI> Allow indexing of more than eight columns
</UL>
<P>
<STRONG>COMMANDS</STRONG>
<UL>
<LI> ALTER TABLE ADD COLUMN to inherited table put column in wrong place [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/inherit">inherit</A>]
<LI> Add ALTER TABLE DROP/ALTER COLUMN feature
<LI> Allow CLUSTER on all tables at once, and improve CLUSTER, loses NOT
<P>
NULL specification on table [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/cluster">cluster</A>]
<LI> Add SIMILAR TO to allow character classes, 'pg_[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/a-c">a-c</A>]%'
<LI> Auto-destroy sequence on DROP of table with SERIAL(Ryan)
<LI> Allow LOCK TABLE tab1, tab2, tab3 so all tables locked in unison
<LI> Allow INSERT/UPDATE of system-generated oid value for a row
<LI> Allow ESCAPE '\' at the end of LIKE for ANSI compliance [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/like">like</A>]
<LI> Rewrite the LIKE handling by rewriting the user string with the
supplied ESCAPE [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/like">like</A>]
<LI> -Move LIKE index optimization handling to the optimizer
<LI> Allow RULE recompilation
<LI> Support UNION/INTERSECT/EXCEPT in sub-selects
<LI> Allow DELETE and UPDATE to use inheritance using tablename*
</UL>
<P>
<STRONG>CLIENTS</STRONG>
<UL>
<LI> Make NULL's come out at the beginning or end depending on the
ORDER BY direction
<LI> Allow flag to control COPY input/output of NULLs
<LI> Update reltuples from COPY command
<LI> Allow psql \copy to allow delimiters
<LI> Add a function to return the last inserted oid, for use in psql scripts
<LI> Allow psql to print nulls as distinct from "" [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/null">null</A>]
</UL>
<P>
<STRONG>EXOTIC FEATURES</STRONG>
<UL>
<LI> Add sql3 recursive unions
<LI> Add the concept of dataspaces
<LI> Add replication of distributed databases
<LI> Allow queries across multiple databases
</UL>
<P>
<STRONG>MISC</STRONG>
<UL>
<LI> Increase identifier length(NAMEDATALEN) if small performance hit
<LI> Allow row re-use without vacuum(Vadim)
<LI> Create a background process for each database that runs while
database is idle, finding superceeded rows, gathering stats and vacuuming
<LI> Add UNIQUE capability to non-btree indexes
<LI> -Certain indexes will not shrink, i.e. oid indexes with many inserts
<LI> Restore unused oid's on backend exit if no one else has gotten oids
<LI> Have UPDATE/DELETE clean out indexes
<LI> Allow WHERE restriction on ctid
<LI> Allow cursors to be DECLAREd/OPENed/CLOSEed outside transactions
<LI> Allow PQrequestCancel() to terminate when in waiting-for-lock state
<LI> -Transaction log, so re-do log can be on a separate disk by
with after-row images(Vadim) [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/logging">logging</A>]
<LI> Populate backend status area and write program to dump status data
<LI> Make oid use unsigned int more reliably, pg_atoi()
<LI> Allow subqueries in target list
<LI> Put sort files, large objects in their own directory
<LI> Do autocommit so always in a transaction block(?)
<LI> Show location of syntax error in query [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/yacc">yacc</A>]
<LI> Redesign the function call interface to handle NULLs better [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/function">function</A>]
<LI> Document/trigger/rule so changes to pg<U>shadow recreate pg</U>pwd [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/pg_shadow">pg_shadow</A>]
<LI> Missing optimizer selectivities for date, r-tree, etc. [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/optimizer">optimizer</A>]
<LI> Overhaul mdmgr/smgr to fix double unlinking and double opens, cleanup
<LI> Overhaul bufmgr/lockmgr/transaction manager
<LI> Add PL/Perl(Mark Hollomon)
<LI> Make postgres user have a password by default
<LI> Add configure test to check for C++ need for *.h and namespaces
<LI> Allow BLCKSZ <= 64k, not <= 32k
<LI> redesign UNION structures to have separarate target lists
<LI> Allow multi-level query trees for INSERT INTO ... SELECT
</UL>
<H2><A NAME="section-1.3">PERFORMANCE</A></H2>
<P>
<STRONG>FSYNC</STRONG>
<UL>
<LI> -Allow transaction commits with rollback with no-fsync performance [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/fsync">fsync</A>]
<LI> -Prevent fsync in SELECT-only queries
</UL>
<P>
<STRONG>INDEXES</STRONG>
<UL>
<LI> Use indexes in ORDER BY for restrictive data sets, min(), max()
<LI> Pull requested data directly from indexes, bypassing heap data
<LI> Use index to restrict rows returned by multi-key index when used with
non-consecutive keys or OR clauses, so fewer heap accesses
<LI> -Convert function(constant) into a constant for index use
<LI> Allow LIMIT ability on single-table queries that have no ORDER BY to use
a matching index [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/limit">limit</A>]
<LI> Improve LIMIT processing by using index to limit rows processed [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/limit">limit</A>]
<LI> Have optimizer take LIMIT into account when considering index scans [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/limit">limit</A>]
<LI> Make index creation use psort code, because it is now faster(Vadim)
<LI> Allow creation of sort temp tables > 1 Gig
<LI> Create more system table indexes for faster cache lookups
<LI> fix indexscan() so it does leak memory by not requiring caller to free
<LI> Improve <U>bt</U>binsrch() to handle equal keys better, remove <U>bt</U>firsteq()(Tom)
<LI> Allow SELECT * FROM tab WHERE int2col = 4 use int2col index, int8,
float4, numeric/decimal too [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/optimizer">optimizer</A>]
<LI> -Allow optimizer to prefer plans that match ORDER BY
</UL>
<P>
<STRONG>CACHE</STRONG>
<UL>
<LI> Cache most recent query plan(s) [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/prepare">prepare</A>]
<LI> Shared catalog cache, reduce lseek()'s by caching table size in shared area
<LI> elog() flushes cache, try invalidating just entries from current xact,
perhaps using invalidation cache
</UL>
<P>
<STRONG>MISC</STRONG>
<UL>
<LI> Allow compression of log and meta data
<LI> Allow char() not to use variable-sized header to reduce disk size
<LI> Do async I/O to do better read-ahead of data
<LI> -Fix memory exhaustion when using many OR's [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/cnfify">cnfify</A>]
<LI> Get faster regex() code from Henry Spencer <<A HREF="mailto:henry@zoo.utoronto.ca">henry@zoo.utoronto.ca</A>>
when it is available
<LI> Use mmap() rather than SYSV shared memory(?)
<LI> -Process const = const parts of OR clause in separate pass
<LI> Make oid use oidin/oidout not int4in/int4out in pg_type.h
<LI> Improve Subplan list handling
<LI> Allow Subplans to use efficient joins(hash, merge) with upper variable
[<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/subquery">subquery</A>]
<LI> use fmgr_info()/fmgr_faddr() instead of fmgr() calls in high-traffic
places, like GROUP BY, UNIQUE, index processing, etc.
<LI> improve dynamic memory allocation by introducing tuple-context memory
allocation [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/memory">memory</A>]
<LI> fix memory leak in cache code when non-existant table is referenced
<LI> In WHERE tab1.x=3 AND tab1.x=tab2.y, add tab2.y=3
<LI> pass atttypmod through parser in more cases [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/atttypmod">atttypmod</A>]
<LI> remove duplicate type in/out functions for disk and net
<LI> Allow persistent backends [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/persistent">persistent</A>]
<LI> Misc [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/performance">performance</A>]
</UL>
<H2><A NAME="section-1.4">SOURCE CODE</A></H2>
<UL>
<LI> Add use of 'const' for varibles in source tree
<LI> Fix C optimizer problem where fmgr_ptr calls return different types [<A HREF="http://www.postgresql.org/docs/pgsql/doc/TODO.detail/alpha">alpha</A>]
<LI> -Add needed includes and removed unneeded include files(Bruce)
<LI> Make configure --enable-debug add -g on compile line
<LI> Does Mariposa source contain any other bug fixes?
<LI> Remove SET KSQO option if OR processing is improved(Tom)
</UL>
<HR>
<H3><A NAME="section-1.4.1">Developers who have claimed items are:</A></H3>
<UL>
<LI> Billy is Billy G. Allie <<A HREF="mailto:Bill.Allie@mug.org">Bill.Allie@mug.org</A>>
<LI> Brook is Brook Milligan <<A HREF="mailto:brook@trillium.NMSU.Edu">brook@trillium.NMSU.Edu</A>>
<LI> Bruce is Bruce Momjian<<A HREF="mailto:maillist@candle.pha.pa.us">maillist@candle.pha.pa.us</A>>
<LI> Bryan is Bryan Henderson<<A HREF="mailto:bryanh@giraffe.netgate.net">bryanh@giraffe.netgate.net</A>>
<LI> D'Arcy is D'Arcy J.M. Cain <<A HREF="mailto:darcy@druid.net">darcy@druid.net</A>>
<LI> David is David Hartwig <<A HREF="mailto:daveh@insightdist.com">daveh@insightdist.com</A>>
<LI> Edmund is Edmund Mergl <<A HREF="mailto:E.Mergl@bawue.de">E.Mergl@bawue.de</A>>
<LI> Goran is Goran Thyni <<A HREF="mailto:goran@kyla.kiruna.se">goran@kyla.kiruna.se</A>>
<LI> Hiroshi is Hiroshi Inoue<<A HREF="mailto:Inoue@tpf.co.jp">Inoue@tpf.co.jp</A>>
<LI> Jan is Jan Wieck <<A HREF="mailto:wieck@sapserv.debis.de">wieck@sapserv.debis.de</A>>
<LI> Marc is Marc Fournier <<A HREF="mailto:scrappy@hub.org">scrappy@hub.org</A>>
<LI> Massimo Dal Zotto <<A HREF="mailto:dz@cs.unitn.it">dz@cs.unitn.it</A>>
<LI> Michael is Michael Meskes <<A HREF="mailto:meskes@postgresql.org">meskes@postgresql.org</A>>
<LI> Oleg is Oleg Bartunov <<A HREF="mailto:oleg@sai.msu.su">oleg@sai.msu.su</A>>
<LI> Peter is Peter T Mount <<A HREF="mailto:peter@retep.org.uk">peter@retep.org.uk</A>>
<LI> Ryan is Ryan Bradetich <<A HREF="mailto:rbrad@hpb50023.boi.hp.com">rbrad@hpb50023.boi.hp.com</A>>
<LI> Stefan Simkovics <<A HREF="mailto:ssimkovi@rainbow.studorg.tuwien.ac.at">ssimkovi@rainbow.studorg.tuwien.ac.at</A>>
<LI> Tatsuo is Tatsuo Ishii <<A HREF="mailto:t-ishii@sra.co.jp">t-ishii@sra.co.jp</A>>
<LI> Tom is Tom Lane <<A HREF="mailto:tgl@sss.pgh.pa.us">tgl@sss.pgh.pa.us</A>>
<LI> Thomas is Thomas Lockhart <<A HREF="mailto:lockhart@alumni.caltech.edu">lockhart@alumni.caltech.edu</A>>
<LI> TomH is Tom I Helbekkmo <<A HREF="mailto:tih@Hamartun.Priv.NO">tih@Hamartun.Priv.NO</A>>
<LI> Vadim is "Vadim B. Mikheev" <<A HREF="mailto:vadim@krs.ru">vadim@krs.ru</A>>
</UL>
</BODY>
</HTML>
functions. One problem that I have encountered with the function
manager is that it does not allow the user to define type conversion
functions that convert between user types. For instance if mytype1,
mytype2, and mytype3 are three Postgresql user types, and if I wish to
define Postgresql conversion functions like
I run into problems, because the Postgresql dynamic loader would look
for a single link symbol, mytype3, for both pieces of object code. If
I just change the name of one of the Postgresql functions (to make the
symbols distinct), the automatic type conversion that Postgresql uses,
for example, when matching operators to arguments no longer finds the
type conversion function.
The solution that I propose, and have implemented in the attatched
patch extends the CREATE FUNCTION syntax as follows. In the first case
above I use the link symbol mytype2_to_mytype3 for the link object
that implements the first conversion function, and define the
Postgresql operator with the following syntax
The patch includes changes to the parser to include the altered
syntax, changes to the ProcedureStmt node in nodes/parsenodes.h,
changes to commands/define.c to handle the extra information in the AS
clause, and changes to utils/fmgr/dfmgr.c that alter the way that the
dynamic loader figures out what link symbol to use. I store the
string for the link symbol in the prosrc text attribute of the pg_proc
table which is currently unused in rows that reference dynamically
loaded
functions.
Bernie Frankpitt
now that sequence names are properly quoted for field defaults, mixed
case sequence names are generated. These are properly quoted in the
CREATE SEQUENCE lines, but not in the SELECT nextval lines, as per
below:
CREATE SEQUENCE "Teams_TeamID_seq" start 10 increment 1 maxvalue
2147483647 minvalue 1 cache 1 ;
SELECT nextval ('Teams_TeamID_seq');
This needs to be:
SELECT nextval ('"Teams_TeamID_seq"');
Patch included below.
--
Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu>
Bruce Momjian