17. Improving Database and Query Performance : Locking and Concurrency Issues : Overflow and Performance : Storage Structure and Overflow
 
Share this page                  
Storage Structure and Overflow
Here are overflow considerations for each storage structure:
Heap--Heap tables are created as one main page with an overflow chain. There is no overflow management.
Hash--Overflow pages occur in a newly modified table if the key is repetitive; this is normal but undesirable. Check a freshly modified table. If there is overflow, consider using ISAM instead.
ISAM--ISAM has a fixed index that can cause long overflow chains. Modify frequently or use B-tree for a non-static table. Use heap structure for large bulk updates and modify back to ISAM to avoid update performance problems.
B-tree--No overflow if there are no duplicate keys, so consider making keys unique. Overflow occurs only at the leaf level and only when 2K pages are used. Use the Shrink B-tree Index option to reorganize it. Use heap structure for bulk loads, modify to B-tree.