ISV HDAM Load Utility Query
search cancel

ISV HDAM Load Utility Query


Article ID: 21969


Updated On:


Database Analyzer (IMS Tools) Mainframe Configuration Manager for IMS for z/OS IMS TOOLS - MISC Compress Data Compression (IMS Tools) Database Analyzer for IMS for z/OS Database Copier for IMS for z/OS Database Organizer for IMS for z/OS Mainframe Extended Terminal Manager (IMS Tools) High Performance Recovery for IMS for z/OS Database Organizer (IMS Tools) Mainframe Program Restart Manager for IMS for z/OS Secondary Index Builder for IMS for z/OS Secondary Index for IMS for z/OS DATABASE MANAGEMENT SOLUTIONS FOR IMS FOR Z/OS


When using an ISV IMS DB "load" utility to load an IMS HDAM database, a "UTILITY DOES NOT SUPPORT RANDOM LOAD OF HDAM DATABASES"
message is manifested and the "load" utility abends with a U4094. "HDAM root segments must be presented to the utility in randomized relative
byte address (RBA), root anchor point (RAP), and key sequence."

Upon further investigation, the "and key sequence" portion appears to be that which is being violated. HP ES IMS DBA Delivery is attempting to change the key with a reload exit.

Received the following regarding the IMS perspective:
"In an HDAM or a PHDAM database, the order in which you load database records does not matter. The user randomizing module determines where each root is stored. However, as with all types of databases, when the database is loaded, all dependents of a root must be loaded in hierarchical sequence following the root". the Database Management Solutions for IMS have the same restriction as the "other" ISV?


No, our DBO tool does NOT have this restriction - it will load the input records, no matter if they are sorted in physical randomizer sequence or not.

BUT (and this seems the reason why the vendor's product only allows loading records in randomizer sequence) : If the input is NOT in randomizer sequence, the load process needs much more time (because of "jumping" around within the dataset) compared with loading records which are in randomizer sequence (and thus can be loaded in physical sequence without jumping around). Therefore for huge datasets, loading data not sorted in randomizer sequence is almost impossible.

BUT: with our DBO tool this problem does NOT occur, because in case of input data which are not in randomizer sequence, you can use an execution option "HDSORT=Y" which then re-sorts the input into randomizer sequence before loading them, so that a fast load process always is ensured.