The partitioning by column option is new for me. It’s a very interesting feature but i wonder in which business case i could use this. Who has used this feature and why ?
I think you're correct in that this is not used often.
This is a performance feature (typically reducing IO's) and so the reason for using it is to make queries run faster. There is no new functionality offered by this feature. So that is the 'business case'.
When to use it? IMHO you need to have a group of (usually end user) queries that use an overlapping subset of the columns in a table. Those columns would be put into a one or more containers and 'other' columns would be put into separate containers.
The advice is that tables which are 'CP' (column partitioned) should not be subject to partial maintenance, i.e. insert/update/delete a portion of the table, instead they should be completely rebuilt each time.
I've looked at using this a couple of times with customers, but you have to know the queries very very well to be able to make a decision about which containers each column is put into. The customer decided that (at the time) they didn't.
Not sure how much that helps but those are some thoughts.
Thanks for the feedback. It was certainly helpful.
We should indeed clarify if we need certain ‘containers’