When an update transaction commits, the database server B-tree scanner threads remove deleted index entries and, if necessary, rebalance the index nodes. The B-tree scanner automatically determines which index items are to be deleted, based on a priority or hot list. The hot list keeps track of how many times an index item caused the server to do extra work. However, depending on your application (in particular, the order in which it adds and deletes keys from the index), the structure of an index can become inefficient.
Use the oncheck -pT command to determine the amount of free space in each index page. If your table has relatively low update activity and a large amount of free space exists, you might want to drop and re-create the index with a larger value for FILLFACTOR to make the unused disk space available.
For more information on how the database server maintains an index tree, see the chapter on disk structure and storage in the IBM Informix: Dynamic Server Administrator's Reference.
Enterprise Edition Home | Express Edition Home | [ Top of Page | Previous Page | Next Page | Contents | Index ]