In this section: |
This section discusses retrieval performance for multi-table Master Files and dynamic joins of three or more tables. It also explains the role of unique cross-referenced tables.
Master Files and JOIN commands define data retrieval paths. The adapter queries only those RDBMS tables necessary for the report. These include tables containing fields specified in the request plus any connecting tables needed to construct a retrieval plan (subtree). The retrieval sequence of a subtree is top to bottom, left to right.
Note: A TABLE request that references a dynamically joined structure generates SQL join predicates for all segments in the subtree that starts from the root segment. Multi-table Master Files do not necessarily generate these predicates. In a multi-table structure, the subtree effectively begins with the highest referenced segment. This difference may cause identical TABLE requests to produce different reports when run against a dynamic join structure and a multi-table Master File that represent the same tree structure.
FOCUS treats unique tables as extensions of their hosts. In cases where the host has both unique and non-unique cross-referenced tables, FOCUS always retrieves the unique cross-referenced tables first.
To display retrieval paths, use the CHECK FILE command with the RETRIEVE option. (See CHECK FILE).
Fields specified in sort phrases (BY and ACROSS) and screening tests (IF and WHERE) must lie on the same retrieval path as the other requested fields. That is, the table containing the BY field or column being screened must precede or follow other tables referenced in the request. A field that is not on the same path generates FOCUS error message FOC029.
| Information Builders |