Forum Discussion
PowerQuery column or DAX group?
When should I lean on a specific solution instead of the other one?
Which one is the least computationally heavy?
Hi Anonymous ,
Normally, the column performance in Power Query will be better than the calculated column in DAX.
Because a column with better compression is smaller in memory and usually provides better performance levels. This is important in filter, group, and aggregation operations involving the column. The compression of a DAX calculated column might be lower than that of a Power Query computed column.
So if you can do it in Power Query/M, you should (except when you are adding a column to a table that references a column in a different table).
If a calculated column or a measure will work, use a measure.Best Regards,
Stephen Tao
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
2 Replies
Please see this article. However, you should use a measure as your first option unless you definitely need a column (for an axis, legend, etc.). In practice, there isn't a big difference for most models but I typically do it in the query if I can (so it is compressed upon loading).
Comparing DAX calculated columns with Power Query computed columns - SQLBI
Pat
Hi Anonymous ,
Normally, the column performance in Power Query will be better than the calculated column in DAX.
Because a column with better compression is smaller in memory and usually provides better performance levels. This is important in filter, group, and aggregation operations involving the column. The compression of a DAX calculated column might be lower than that of a Power Query computed column.
So if you can do it in Power Query/M, you should (except when you are adding a column to a table that references a column in a different table).
If a calculated column or a measure will work, use a measure.Best Regards,
Stephen Tao
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
