So, I've spent a little bit of time on this
I am doing a little bit of manipulation on staging (also comparing to what exists) before moving it live - so I am finding staging indexes useful
Here I do all this in an interim live table and partition switch live as a final step
It's a full load every time as it's ever changing 3rd party data
So, (in with the testing theme that you have all suggested above), some tables are faster and some slower
I've asked the question as I was advised not to drop and recreate clustered indexes every time. To either leave them on or not have them at all (as pointed out in your answers). Again, downstream, this slows some of my tranformation actions down.
Thing is, I have a couple of examples where leaving them on is a lot slower
Is there an appraoch to ordering first that's required. Must admit I've tried alsorts here and read a lot of different articles with no real success.
Seems to me as though dropping and recreating an index sometimes is the best way.
Are there any negatices to that approach?