EIM - Best Practice

1. ONLY BASE COLUMNS or IGNORE BASE COLUMNS parameters
List only those columns that are relevant for a particular EIM task.

2. ONLY BASE TABLES or IGNORE BASE TABLES parameters
List only those tables that are relevant for a particular EIM task.

3. USE INDEX HINTS Parameters usage
If FALSE, EIM does not generates hints during processing
EIM processing should be tested with both settings (TRUE and FALSE) to determine which provides better performance.

4. Additional Indexes on EIM tables
Sometimes it is recommended to create additional indexes on EIM tables to improve performance of time-consuming SQL statements.

5. Separate Insert and Update Statements
Always better to separate the Inserts and Updates in different batches
EIM has to perform additional processing to determine whether to insert or update

7. Controlling size of batches
Siebel recommends to use a batch size no more than 5000 rows.
Using batch ranges (x-y) allows us to run with smaller batch size and avoid the startup overhead on each batch.

8. USING SYNONYMS parameter usage
When set to FALSE, it saves processing time because queries that look up synonyms are not used (for Account load) Should not be set to FALSE if multiple addresses are used for Accounts.

9. Transaction logging
When set to FALSE during initial load, reduces transaction activity to the Siebel docking tables.

10. Parallel execution of EIM jobs
Run mutually exclusive EIM processes concurrently.

11. Optimizing SQL
Running EIM job with the following flag settings:
Error Flags: 1
SQL Trace Flags: 8
Trace Flags: 3

12. UPDATE PRIMARY KEYS and PRIMARY KEYS ONLY parameter usage
UPDATE PRIMARY KEYS and PRIMARY KEYS ONLY parameters should be avoided in IFB file.

13. Regular Table maintenance (GATHER TABLE STATS)
Running gather stats would help faster processing of batches.

14. Clear EIM tables
Delete old batch records before running fresh EIM task and also Truncate if possible.