Tuning IBM Content Cortex queries
As the IBM Content Cortex
administrator or database administrator, you determine what constitutes the optimal set of database
indexes for your system and needs. You do so by analyzing the queries that are typically run by your
Content Cortex applications.
Content Platform Engine does not come installed with a default set of
database indexes that supports high performance searches for all application designs. No such single
set of database indexes is possible for the following reasons:
| Reason | Comment |
|---|---|
| Powerful query capability | Content Platform Engine supports customizable queries against classes and properties. These queries can be constructed with various operators and combinations of search conditions, joins, subqueries, and ordering. Also, queries can include the paging of results. |
| Performance trade-offs | Some trade-offs are necessary to provide high performance overall in terms of both searching for and creating objects. Unnecessary indexes can degrade performance when objects are added or updated. |
Important: A query can be either a relational query or a content-based retrieval (CBR)
query. A CBR query consists of two main constituent searches: a full-text search and a database
search. For the final join of these two searches, the indexing solutions that are presented here
apply. Also, the guidelines here apply generally to CBR queries except for the full-text search
aspect of these queries. For more information, see CBR
Queries.