c-treeACE SQL imposed a restriction of only one instance of an index name per database. For example, an inventory database with a customer table and an invoice table could not have an index on each table named “date.” This was loosely implied by the fact that the database engine did not create index files with a unique name per table. This enhancement removes this restriction by pre-pending the table name to the index file name, thus removing the index file name collision problem.
This support could cause unexpected problems in applications that utilize a c-treeDB API and have hard-coded index names for tables created using c-treeACE SQL. Therefore, a configuration keyword has been added to revert to the original behavior.
SQL_OPTION NO_DUP_IDXNAME
With this keyword present in the ctsrvr.cfg configuration file, index naming is as before and it is not possible to create indexes with the same name even if belonging to different tables.