How to use Power Query (M) efficiently

[robarivas]

I created "QueryZ" (using Append) that establishes a population of accounts coming from 2 tables. I want to use those accounts as a filter against other tables. So I do a Table.Join against some table (call it TableX) hoping the load process will only scan a limited number of rows from TableX. However, Power BI still seems to try to scan the entirety of TableX during the load process. Is there a best practice of how to do mergeing in Power Query (or otherwise leveraging the accounts from Query1) to prevent the load process from scanning the full TableX.

This issue is important to me because I happen to be dealing with massive tables so I need the tool to not evaluate excess rows unneccesarily.

[ImkeF]

You might experience this bug in query folding: http://www.thebiccountant.com/2017/07/20/sql-query-folding-bug-still-alive-sucking-powerbi-powerquer...

[stretcharm]

There are some tips on this page.

https://exceleratorbi.com.au/best-practices-power-pivot-power-query-power-bi/

Also you can turn on tracing from the setting menu which may help you see what it's doing and if any changes are making it better or worse.

There are some buffer functions that may help.

https://blog.crossjoin.co.uk/2015/05/05/improving-power-query-calculation-performance-with-list-buff...

Some posts that may have useful details

https://community.powerbi.com/t5/Desktop/How-to-Improve-Query-Reference-performance-for-large-tables...

http://community.powerbi.com/t5/Desktop/merging-tables-very-slow-performance/td-p/122730