Bookshelf Home | Contents | Index | Search | PDF

Siebel Enterprise Integration Manager Administration Guide > EIM Performance Tuning > Creating Proper Statistics on EIM Tables >

Dropping Indexes in Initial Runs

---

Typically, the EIM initial load is a very database-intensive process. Each row that is inserted into the base table requires modifications on the data page and the index pages of all the affected indexes. However, most of these indexes are never used during an EIM run. Index maintenance is a very time-consuming process for most database managers and should be avoided as much as possible. Therefore, the goal is to determine any indexes that are unnecessary for EIM and that can be dropped for the durations of the EIM run. You can create these indexes later in batch mode by using parallel execution strategies available for the respective database platform. Using this approach can save a significant amount of time.

NOTE:  Under normal operations, using parallel execution strategies is not recommended.

• Target Table Indexing Strategy. For a target base table (such as S_ORG_EXT) you only need to use the Primary Index (Px for example P1), and the Unique Indexes (Ux for example U1), and then drop the remaining indexes for the duration of the EIM import. Past experience has determined that the Fx and Mx indexes can be dropped after an extensive SQL analysis of sample EIM runs.
• Non-target Table indexing Strategy. For child tables (such as S_ADDR_ORG) you only need to use the Primary Index (Px), the Unique Indexes (Ux), and the Foreign Key Indexes (needed for setting primary foreign keys in the parent table). Past experience has determined that the Fx and Mx indexes can be dropped after an extensive SQL analysis of sample EIM runs.

NOTE:  Testing should always be performed when dropping indexes (or adding indexes) to make sure that expected results are achieved.

Bookshelf Home | Contents | Index | Search | PDF