Sequence cache definition

On our radar
It would be nice to be able to define the cache used by the sequences created by the platform. It could be configured in the server configuration, it would not be necessary to configure it per usage.

This way, we could cache the sequences created in Oracle databases to mimify the autonumber feature it is lacking, to improve performance.
Created on 2 May 2011
Comments (3)
25 Aug 2015
Really good idea Andre.

In fact I have this performance issue now. I have a timer that loads a huge volume of data from a CSV file. 60% of the time is consumed by a sequence's next_val operation.

I just post a thread in the forum about this: https://www.outsystems.com/forums/discussion/15786/how-about-changing-cache-definition-of-sequences/

26 Aug 2015
I do not disagree with it, but @joao, why not use bulkimport for enormous amounts of data?

27 Aug 2015
J., I need to evaluate the data while importing. Some data must be validated in other tables, if it's not there we generate a report, etc, etc...