New in version 2.20.16
November 13th, 2014
- Added utility function blobsize for retrieving blob size information.
New in version 2.20.6 (February 25th, 2014)
- Added fix in CegoFunction::evalFieldValue for SUBSTR case to avoid stringbuf exceed exception. start and len parameters are set to string boundaries in case of invalid user input.
- Added csmode and qescmode flags for cego xml initialization Exception message fix in CegoFactor::evalFieldValue.
New in version 2.20.4 (February 15th, 2014)
- Added hitrate calculation to buffer pool information.
- Added transitive role removal for all users, if a role is dropped.
- Added many improvements to curses based administration frontent ( cgadm ) Now tableset, user and role administration can be done supported by an curses based GUI. Null value check added for "alter table" command, so in case of modification of an attribute to be not null, the table entries are checked for it (new method CegoTableManager::checkNullValue)
- Cleanup in CegoObjectManager ( introduced method getHashPageId) Fixes for table altering : In case of renaming table attributes, the corresponding key entries for index, btrees, fkey and checks have not been modified.
New in version 2.20.3 (February 1st, 2014)
- Fixes added in CegoBufferPool::writeAndRemoveTabSet and CegoBufferPool::writeCheckPoint so just pages are written, which are occupied. Otherwise, this may lead to system crash, if pages which have been released ( marked as NOT_OCCIPIED ) and written to disk.
New in version 2.20.2 (January 29th, 2014)
- Fixed a problem in CegoFileHandler::allocatePage If a page is allocated with doAppend == false, the append pointer is reset. Otherwise, freed pages are never reallocated again and sustained create/drop operations may lead to page exceed.
New in version 2.20.1 (January 27th, 2014)
- Fix for btree objects. In some cases, bufferUnfix operations have not been done correcty, which might result in corrupted btree objects.
New in version 2.20.0 (January 21st, 2014)
- The CegoBTreeManager::deleteBTree method has been redesigned. In case of non-unique index objects, it might happen, that the corresponding index entry was not found ( if the value is spreaded over several leaf nodes ) This problem is fixed with the new method implementation
- Added relevance attribute value for btree object. This allows to make a better decision about the relevance of a btree object which results in an optimized query execution plan.
New in version 2.19.14 (January 6th, 2014)
- Functional expansion for aggregation queries to support aggregation in functions e.g. select f1(sum(a)) from t1.
New in version 2.19.13 (November 23rd, 2013)
- Further corrections for CegoDatabaseManager::useObject method numTries was not increased correctly.
- Optimization for CegoTableManager::deleteDataTable The index evaluation was not correct, so the faster deletion strategy for tables without index objects was not used.