Reputation: 129
I am trying to connect my Hadoop clusters (Cloudera) from Polybase 2016. When I am querying Hadoop clusters from Polybase, the performance I am getting is very slow. It is in fact, 50 times slower than local SQL Server tables on same data size. Did any one has worked on to improve query performance on Polybase?
Regards,
Ritu Ranjan
Upvotes: 0
Views: 759
Reputation: 810
This is expected behaviour as Polybase generates Map Reduce code. Map Reduce at this stage is considered legacy technology. Polybase also has various other pushdown limitations, e.g. it doesn't push down joins into Hadoop.
Upvotes: 2