I have few questions about backup strategy as it is taking a long time to backup full data.
Is incremental backup possible using TARA jobs?
Is it a must to have a primary partitioned index on a table to have incremental backup only based on that specific PPI column ?
Can we not use a simple date column like update_record_date as a partitioned column only for the backup purpose while there is no PPI defined on update_record_date ?
if we have a text column partitioned how can we use it for incremental backup ?
what will be the strategy to restore such backup based on text based partition ?
I will really appreciate your support.
No, ARC / TARA does not support true incremental backup.
Yes, PPI is necessary to back up only a subset of rows from a table.
No, you can't use a non-PPI column. (You could separately use SQL to extract a subset of data to a separate table, and back up that table.)
There's nothing inherently special about date versus character PPI in this context.
See the Teradata Archive/Recovery Utility Reference manual discussion on considerations and limitations of backing up and restoring selected partitions.
Yes, in principle this should work. Be sure you review and understand the limitations / dangers of partition based backup and restore.
I would put backups of selected partitions in a separate job, just to make it a bit less likely that someone would accidentally restore the partial backup to the table (and wipe out the rest of the data). And you likely need extended retention on the PPI backups anyway (as well as the full backups of those tables, done less often).
Minor point: Note that to "restore PPI based backup in different tables" you will need to use the COPY command instead of RESTORE.