Forum Discussion
Difference between writing a native query and applying a filter in M
AFAIK it's basically as Greg_Deckler said.
If you're writing custom SQL, you have to be 100% sure of your SQL skills and perform all the transformations in the SQL script, because it will break query folding and you won't be able to fold any step done in the PQ interface.
If your data source supports query folding and you're performing very simple transformations such as filtering columns, these should be folded and you shouldn't notice any difference in performance.
Anonymous My understanding is that if the query can be folded then it should essentially result in being the same thing both ways. However, if it cannot be folded, then Power BI would need to load all of the data and then filter it down which could be a significant performance it. You should be able to test this quite easily as SQL connector allows you to put SQL code into the base query.
AFAIK it's basically as Greg_Deckler said.
If you're writing custom SQL, you have to be 100% sure of your SQL skills and perform all the transformations in the SQL script, because it will break query folding and you won't be able to fold any step done in the PQ interface.
If your data source supports query folding and you're performing very simple transformations such as filtering columns, these should be folded and you shouldn't notice any difference in performance.