SQL!$@w$0ME - Monday, May 22, 2017 12:35 PM
To be honest, I probably wouldn't bother. With the data distributed, possibly partitioned and spread out across all the nodes, fragmentation probably isn't a huge concern.
You MUST update statistics manually (and create them manually too). The control node does not create or update statistics automatically and it having correct statistics is critically important to get good distributed plans.
Are you using mostly columnstore indexes, as is the recommendation I've seen a lot, or rowstore clustered indexes?
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability