Forum Discussion
Regression Formula
Just check your syntax. You can try this formatter: www.daxformatter.com.
The line that is incorrect:
VAR KnownX = ALLSELECTED ( 'All_Actual_Sales (3)', 'All_Actual_Sales (3)'[LN(x)])
The error the formatter reports is the comma ofter the first occurence of 'All_Actual_Sales (3)'. Please check the syntax of ALLSELECTED. You cannot put a table and then, after a comma, a column under ALLSELECTED. You can put in there either a table name, or a (set of) column(s) or nothing at all. You are trying to mix things incorrectly.
Thank you, so what expression would I use if I a simply trying to say, KnownX is column LN(x) for each row?Distinct?
Can't give any advice since I don't know what the model looks like.
The dificulty is that the columns I am using are from 2 diffrent tables. I have tried to use power query to merge the tables but becuase they are so large, power query consitently crashes. (3 m rows). So Im trying to write a regression formula using "LN(x)" which is a calculated column and "cost % column" which is another calculated column.
Power Query can be written in an optimized manner as well, just like DAX. Not all DAX is performant as well as not all M is. One has to investigate how to write it to make it fast. 3M rows does not seem to be an overly excessive figure for M not to be able to handle this gracefully and quickly. I think your code is just.... suboptimal. Please, for any M optimizations consult the blog of Chris Webb, Chris Webb's BI Blog: Power Query Chris Webb's BI Blog (crossjoin.co.uk). Or, alternatively, you could ask a question about it on the PBI forum dedicated to Power Query/M.