That last measure looks great! I figured that's where you were heading towards, but I couldn't quite parse it out the way you had it first written.
Always use variables. If you evaluate the same expression more than once, make it a variable.
2 other pieces of advice:
FILTER() is an iterator, so it will look at EVERY row in the table in the current filter context, and include the row if it passes the 3 criteria that you set. I would recommend the following code for VAR FilteredTable:
VAR CurrentBusUnit = SELECTEDVALUE ( 'MTHLY'[IBU], "" ) VAR FilterTable = CALCULATETABLE ( 'CUS_RANGE', 'CUS_RANGE'[BUS_UNIT] = CurrentBusUnit, 'CUS_RANGE'[SUMMARY_TYPE] = "ACCT_NBR", 'CUS_RANGE'[METRIC_NAME] = "APPT_WIN_HR" ) ...rest of the measure
Instead of looking at every row in the current filter context, this expression will evaluate the entire table in the context of the 3 filters. Should take you below 10 seconds.
You may be able to replace that first parameter of CALCULATETABLE() (the 'CUS_RANGE') with a
VALUES( 'CUS_RANGE'[ColumnName] )
if you have a unique identifier column. That will then return only the distinct list of values for that specific column where the other 3 columns are as specified. That single column table will then act as the filter for everything else.
In short, don't use FILTER() on an entire table if you can get around it.
Based on this discussion, I think my DAX code would benefit from optimization. My report takes up to a minute to load visuals after making a slicer selection, and it makes the report unusable. I'd appreciate any suggestions to improve my work!
I have a central fact table, Jobs, and I'm trying to add columns with data from other related fact tables (using the ID field). Item1, Item2, etc. data are spread across a handful of tables and I'm trying to get counts for Fail and Pass for all Items all in one place:
Results = ADDCOLUMNS(Jobs, "Item1_PASS", COUNTAX ( FILTER ('DataTable1', 'DataTable1'[Item1] = "Pass" && 'DataTable1'[ID]=Jobs[ID]), 'DataTable1'[Item1]), "Item1_FAIL", COUNTAX ( FILTER ('DataTable1', 'DataTable1'[Item1] = "Fail" && 'DataTable1'[ID]=Jobs[ID]), 'DataTable1'[Item1]),
My understanding was that I need to create measures (in a separate measure table) that calculate a sum for each of the columns that I'm adding to Jobs:
Item1_FAIL = SUM('Results'[Item1_FAIL])
The end goal is a visual with counts of Pass and Fail for each Item that will work with slicers that use other columns from Jobs, like this:
Fail Count = SWITCH ( FIRSTNONBLANK( 'Item_key'[Item Name], 1), "Full description for Item1.", 'My Measures'[Item1_FAIL],
Again, this report has become very frustrating to use because of the load times - I have over 200 Items to display results for. I imagine I could set up my code in a better way, but I'm not sure what that would be. Thanks in advance for suggestions, and please let me know where I can clarify anything.
@jscottNRG, please post this response in a new thread, and mention me in it.
I think it's best practice to keep a thread to a single issue. Measure Optimization is a broad topic, but this deserves its own thread.
Also, please include a picture of your relationships so that I can see the other dimension tables as well.