Hello All,
I assume people have already started deploying Table Partitioning in 11.6 to Production by now.
a) What kind of performance improvement do we get (big/small)
b) In terms of the 4GL side of things - What would be the application gotchas other than the few that are listed in the documentation with respect to making sure that partitioned fields are assigned during the create.
Thanks for sharing
I do not think there are many implementations of Table Partitioning. I know only one done by PROGRESS Professional Services Team (aka Mike Furgal team)
Look at PowerPoint presentation by Rich Banville at PUG Challenge 2015
#483: Getting the Most out of Table Partitioning
pugchallenge.org/downloads2015.html
Rich is talking about performance improvements towards the end of presentation.
From myself, I can just add that I am using Table Partitioning with Oracle for the last 12 years. There are definitely performance improvements from doing that and I believe same or similar improvements should happen with Progress DB.
One area that is often missed -- ease of Administration. It is so much more easy to administer database that has partitioning on a big table, that that alone is worth doing partitioning. I move, archive, compress, drop Oracle partitions easily.
Also local indexes (index on a partition) are much smaller then global indexes (index on a whole table). Smaller indexes should be faster. Thus reading should be faster.
Also Progress partitioning will allow to rebuild local index online. It is not a problem for Oracle where any index could be rebuild online, but for Progress it is a big step forward. One of many.
There is also presentation by Gus that may be of interest to you