Catalog search performance
The Catalog Search feature uses what is known as a vertical schema search. A vertical schema search searches through tables in which data is stored vertically in name/value pairs, rather than through data stored horizontally in predefined columns. This flexibility impacts performance to some degree because the data is not optimized for searching. The simple catalog search will perform about as fast as a simple text string search and customers will not experience a significant delay between launching a search and obtaining results, except those caused by common Internet traffic constraints.Advanced catalog searches may be considerably slower due to the complex SQL queries that are generated by the underlying data bean. In most cases, the response times are not prohibitive but you may want to alert your
customers that there are factors that can adversely effect performance, such as database size and configuration, number of search attributes, number of results, hardware, and memory.
It is strongly recommended that you try your queries before putting them into a production environment.
Use the following techniques to improve Catalog Search performance:
- Populate your data prior to creating summary tables. The data population rate can be slowed by as much as 90% if summary tables are created in advance. When using large data sets, defer the creation of summary tables until after the data population is completed on the base tables. Larger or additional primary and secondary DB2 log files may be needed to accommodate creation of summary tables in this manner.
- Do not include rich attributes in search constraints or catalog search performance will slow down considerably; especially if you use more than three rich attributes in the search.
- To increase speed, perform searches according to attributes that are stored in a single table, such as list price, rather than searching attributes that are stored in summary tables, such as standard price. Summary table searches take longer because they contain data from multiple tables.
- Avoid searching on large data types (greater than 1KB), such as DB2 LONGVARCHAR.
Related concepts
Catalog search methods
Simple catalog search
Advanced catalog searchRelated tasks
Add catalog search to your siteRelated reference
Catalog search limitations