B-link indexes are a variant of B- and B+- tree indexes in which each index page, non-leaf and leaf, contains the page number of (or a link to) its right sibling. Further, index pages need not appear immediately in a parent page. The primary advantage of B-link indexes is improved concurrency.
Indexes can be declared as either clustered or unclustered. Only one index on a table can be clustered. If you determine that an index should be clustered, you do not need to drop and recreate the index: the clustering characteristic of an index can be removed or added by issuing an ALTER INDEX statement. Clustered indexes may assist performance by allowing the query optimizer to make more accurate decisions about the cost of index scans.
To improve fanout, SQL Anywhere stores a compressed form of each indexed value in which the prefix shared with the immediately preceding value is not stored. To reduce the CPU time when searching within a page, a small look-aside map of complete index keys (subject to data length restrictions) is also stored. In particular, SQL Anywhere indexes efficiently handle index values that are identical (or nearly so), so common prefixes within the indexed values have negligible impact on storage requirements and performance.
Discuss this page in DocCommentXchange.
|Copyright © 2010, iAnywhere Solutions, Inc. - SQL Anywhere 12.0.0|