[Db-derby Wiki] Update of "DerbyToDos" by AndrewMcIntyre

[Apache Wiki <wi...@apache.org>]

Dear Wiki user,

You have subscribed to a wiki page or wiki category on "Db-derby Wiki" for change notification.

The following page has been changed by AndrewMcIntyre:
http://wiki.apache.org/db-derby/DerbyToDos

The comment on the change is:
Dump of the page from the website

New page:
Currently, this is a dump from the old, outdated, and obsolete DerbyToDo page that was on the website. This page serves to maintain items that were on that page that do not have JIRA issues filed until such time as JIRA issues can be filed.

Please update and correct any information you find on this page as necessary. If the item has a JIRA filed and no further action is necessary at this time, feel free to remove that line from the table.

SQL

|| Feature || discussion || JIRA || comments ||
|| ALTER TABLE DROP COLUMN || many || [http://issues.apache.org/jira/browse/DERBY-1489 DERBY-1489] || ||
|| COMMENT ON || [http://mail-archives.eu.apache.org/mod_mbox/db-derby-dev/200409.mbox/%3c4142E381.4020809@infologic.fr%3e derby-dev thread] || || ||
|| User Defined Types || [http://mail-archives.eu.apache.org/mod_mbox/db-derby-dev/200409.mbox/%3c71A4A5D6F4E5634387B896E7651606451AD771@EXCH-BEL4.attachmate.com%3e derby-dev thread]  || [http://issues.apache.org/jira/browse/DERBY-672 DERBY-672] (?) || ||
|| Gist indexing (spatial) || [http://mail-archives.eu.apache.org/mod_mbox/db-derby-dev/200409.mbox/%3c6.1.0.6.0.20040915082545.01e69ae8@marilyn.barrodale.com%3e derby-dev thread]  || ||
|| Localized character sorting || [http://mail-archives.eu.apache.org/mod_mbox/db-derby-dev/200409.mbox/%3c20040920095243.A487C3D86AE@poczta.interia.pl%3e derby-dev thread] and more || related to [http://issues.apache.org/jira/browse/DERBY-533 DERBY-533] || related to NCHAR etc. support ||
|| SQL/XML / XQuery || many || mostly in 10.2? || ||
|| SPECIFIC name for PROCEDUREs and FUNCTIONs || || || SPECIFIC name is a step towards overloading of PROCEDUREs and FUNCTIONs. ||
|| Overloading of PROCEDUREs and FUNCTIONs || [http://mail-archives.apache.org/mod_mbox/db-derby-dev/200410.mbox/%3c41640227.9090502@debrunners.com%3e derby-dev thread]  || || Allows multiple PROCEDUREs and FUNCTIONs with the same name but different parameter types. ||
|| SQL/PSM || [http://wiki.apache.org/db-derby/SqlPsmSupport wiki] more elsewhere || || Would need to be standard based but then would that limit its use as applications that use it might not be able to migrate up to other databases. Need to investigate overlap of standard options with vendor specific procedure language. ||
|| TRUNCATE TABLE || || || Code basically complete in engine, needs to be exposed. Details of rollback behaviour to be clarified. ||
|| National character types (NCHAR etc.) || || [http://issues.apache.org/jira/browse/DERBY-533 DERBY-533] || Code basically complete in engine, need to be exposed and verified with SQL standard. Code could also use cleanup and ensure casts between types is handled correctly. ||
|| Multi-version concurrency locking || || || ||


JDBC

|| Feature || discussion || JIRA || comments ||
|| Updateable ResultSets || || || complete in 10.2? ||
|| Support update methods in java.sql.{Blob,Clob} || || || ? ||
|| JDBC 4.0 support || || || mostly complete in 10.2 pending finalization of JDBC 4 spec ||
|| Query timeout || || DERBY-31 || Complete in 10.2 ||
|| Support java.sql.Statement.cancel() || || || ? ||

Platform Support

|| Feature || discussion || JIRA || comments ||
|| OSGi ee.minimum support || || || ||
|| JDK 5.0 || some discussion of NIO || DERBY-810 || Look into seeing if there are specific features of JDK 5.0 that would benefit Derby.Most likely would split into multiple items.||
|| J2EE 5.0 || || || Place holder - Specific JBDC requirements of J2EE 5.0. Most likely would split into multiple items. ||


Other - Miscellaneous features 

|| Feature || discussion || JIRA || comments ||
|| Dynamic swap of derby.log || || || Better mechanisms for handling the error log with a long running system. E.g. Switch log files every 100k bytes. ||
|| Backup directly to jar file || || || Make the backup code directly create a jar file. Allows easy creation of jar files for 'database-in-a jar 'feature ||
|| Generate run time statistics output in XML format || || || Might reduce footprint by not requiring specific classes for each ResultSet type ||
|| Text search/indexing || many || [http://issues.apache.org/jira/browse/DERBY-472 DERBY-472] || Pluggable mechanism to support text indexing engines, how to make SQL queries standard? ||
|| Migration tool || Google SoC project || [http://issues.apache.org/jira/browse/DERBY-1651 DERBY-1651] || ||
|| In-memory database || many || [http://issues.apache.org/jira/browse/DERBY-646 DERBY-646] || ||
|| Single file per database || || || Rather than a single file per table. ||
|| Property validation || || || Framework and validation for some properties ||
|| JMX Support || Google SoC project || [http://issues.apache.org/jira/browse/DERBY-1387 DERBY-1387] || ||
|| Cache Manager Support || Google SoC project || [http://issues.apache.org/jira/browse/DERBY-1437 DERBY-1437] || ||

Performance/Zero Admin/Footprint

|| Feature || discussion || JIRA || comments ||
|| Query re-writes for queries with constants || [http://mail-archives.apache.org/mod_mbox/db-derby-dev/200409.mbox/%3c413EA784.6030407@jansdal.dk%3e derby-dev thread]  || || ||
|| Smarter query recompilation || || || The current scheme of check every N iterations is simplistic and can be expensive. Also information about which piece of the query was badly estimated is there but not used. ||
|| Autonomic checkpointing timing and log file size || some - need reference || || ||
|| Current checkpoint interval and log file size can be configured, but engine should self-configure. || some - need reference || || ||
|| Use of idle time for background checkpoint || some - need reference (see threads by Raymond Raymond) || || ||
|| Single background thread || || || Currently one background thread is created per database, which becomes an issue when supporting thousands of databases per JVM. ||
|| Improved statement plan cache || || || Make single-use (DDL) statements indicate they don't need to be cached, support single plan regardless of compilation schema if they do not depend on the current schema ||
|| Background space re-claiming || some - need reference || || Allow space to be returned to the operating system by truncating files. Have to move rows around dynamically etc. ||
|| Support super-simple configuration of single configuration parameter - how much total memory to use || || || All caches would re-size themselves based upon the value and their size and number of connections etc. ||
|| Backward scan in index || some - need reference || || ||
|| Implement limt on on-disk database size || some - need reference || || ||