TLDR The short answer is YES - if you have >= 60 million rows, then a poorly maintained column store was 100% slower due to trim fragmentation and 64% slower due to the delta row store with some updates in our tests. Obviously this is just a small test on only 60 million rows and … Continue reading Do we need to maintain Columnstores in SQL Pools ?
One recent realization for myself working with Synapse Dedicated SQL Pools is that while they have auto create stats, there is no auto update stats. So statistics will become stale faster than your bread left in the toaster. Rather than updating every 20% or so like traditional SQL Server. In additon, if you create an … Continue reading Finding Outdated and Missing Stats in Synapse SQL Pools