My current understanding:
- Oracle has index-organized tables and defaults to heap-organized.
- I heard that SQL-Server uses only index-organized tables
I'm especially thinking about solutions for MySQL, Informix and DB2.
MySql has clustered indexes but there seems to become limited control on these indexes.
DB2 has MDC (multi-dimension cluster) which could effectively index organise the table in a number of ways. I have not used at all them however, you can most likely simply have just one dimension MDC which will be the just like a typical clustered index.
Oracle is a little of the discomfort. Before I checked using its IOT implementation requires you produce the table as IOT and also you can't change it out later, that is annoying when you wish to load a lot of information without any indexes for speed after which index it after.
For Informix Dynamic Server (IDS) or Standard Engine (SE), both of them support clustered and non-clustered indexes. IDS uses RSAM and SE uses C-ISAM B-Tree indexing. For, whenever a cluster index is produced, a table's rows are physically purchased within the same order because the index, but as new rows are put into the table, the row is physically placed at EOF (heap).