Hi,
I am a COBOL developer and in a few months my project will start using CA-IDMS as the backend.
I have experience only in using DB2 and IMS as backends and IDMS is new to me
I would be grateful if the IDMS experts can give their opinion on 1 question (and 2 related questions) that has been asked of me.
Question: Which facilitates faster record retrieval:
1) Indexes on tables
OR
2) defining CALCS / RECORDS / SETS on the underlying tables?
Related question: Would the life of IDMS administrators become easier if there were no CALCS+RECORDS+SETs on the IDMS schemas?
Related question: Does having CALCS+RECORDS+SETS cause delay while executing common IDMS utilities like: REORG / LOAD / UNLOAD?
My COBOL team is undergoing a training session and it will take some months for us to start co-ordinating with our client team who are managing IDMS (project is still in SOW negotiation stage)
Meanwhile I have been told to go through the existing COBOL modules and suggest performance improvements.
One thing that struck me is the backend Central Versions have lots of record structures clubbed together in chain linked sets which I understand cause various records to point to each other
Despite the existence of these record sets the COBOL code is still accessing the tables using in-built indexes (.i.e normal select sql statements which cause IDMS to load a specific index during runtime)
I can suggest that "Existing COBOL code be revamped to use the CALC+RECORD+SET objects instead of simple select SQL statements which use only indexes"
But before I suggest I want to make sure that I am not causing extra strain to the backend IDMS admin team
Would appreciate if the IDMS experts can provide some useful pointers on the 3 questions above
Thanks
Infra Application Developer