Nicolas,
The B_SNAPSHOT_<repositoryId> tables mirror the repository records for each referenced repository. These tables are referenced by the "friendly" views that list attributes by restricted names instead of F_<attribute_id>. The EPIM_TPARTNER_SEQUENCE facilitates generating the next sequence number for each of the defined Sequences (originally managed as Trading Partners in the UI for GDSN data pool synchronization but repurposed). Each time a new sequence number is generated, a record is inserted into the corresponding table and that value read. This results in these tables growing over time, but none of the records are needed. It should be safe to truncate their contents as it won't affect the next insert and sequence number generation.
If you are looking to reduce the footprint of the EPIM database, one thing to look at is the Server Properties in the Classic UI under the System feature. This is where the cleanup job is configured and you can specify whether to cleanup different types of records that grow over time and specify how long to keep the older records around. The largest consumer of space would be for the Clean Record History setting. If this is not enabled, or the number of days is very large, then there will likely be a lot of history records. Basically, every time a repository record is changed or deleted, a new history record is inserted into the repository history table. This record contains a complete copy of the data prior to whatever change or delete was performed. In a newly populated repository, altering every record in that repository (e.g., performing a mass edit on a single attribute) will roughly double the storage space used for that repository.
The Clean Saved Sets could also have high storage requirements if not enabled or the number of days is large and the environment has a lot of processing (imports and promotions) as these operations generate temporary saved sets (not visible from the UI) that will remain in the system until they are reclaimed by this process.
If these cleanup settings are enabled or changed, the next time the cleanup process runs, more data will be scrubbed from the database and the files can be subsequently shrunk to use up less space.
If the other tables you mentioned have a lot of data, I suggest you open a Support ticket with the details so descriptions and options regarding their cleanup can be provided.
-Brian
------------------------------
Brian Zupke | Senior Technical Support Engineer
Winshuttle North America | 909-900-9179
------------------------------
Original Message:
Sent: 10-10-2022 10:03
From: Nicolas Carrillo
Subject: Cleaning my EPIM database
Hello Everyone, from a couple of months ago, we realized the need to clean our EPIM Database
- (SNAPSHOT, SNAP_VALIDATION, EPIM TPARTNER SEQUENCE, PLT, VALIDATION_TMP) - TABLES
- (EPIM_REP, EPIMV) - VIEWS
One question while talking about it was:
- what do these tables mean?
- are these required on a daily basis?
- are EW developers allowed to delete this/ should we contact a DBA?
- what tables should we keep (f.e. the last snapshot table)?
Thanks for your time reading me, community
------------------------------
Nicolas Carrillo | Product Data Analyst
Steelcase, Inc | +52 (81) 8248 2500
------------------------------