Forum Discussion
Query folding creates unusable queries in merge
when joining large tables via "Merge Queries"Try to avoid using merges/joins in Power Query as much as possible. They are very expensive in a couple of areas. Instead, let the data model do the work for you (unless you have to use fuzzy joins in which case you're doomed. I mean in which case you need to think hard about cardinality and order of joins.)
Thanks for your reply. We do try to avoid them as much as possible, however currently there are situations for us where (IMHO) there's no alternative.
Also I assumed that it's actually less expensive (on the Power BI side) to do it this way, since the amount of data that is imported is sized down beforehand?
Before merging you should reduce the column count to the smallest viable amount. I believe the data has to be sized down. You could try to unpivot other columns that are not the merge column. I have always tried to Remove columns when it's possible (and they are not needed).