Forum Discussion
Compare two measures to evaluate dimension status
- 3 years ago
Hi b2wise ,
Try the following code:
Fix = VAR temp_table = CALCULATETABLE ( GROUPBY ( SUMMARIZE ( 'Table', 'Table'[Region], 'Table'[Item] "TEST", [Sold Out ERP M] = [Sold Out WMS] ), 'Table'[Item], [TEST] ), REMOVEFILTERS ( 'Table'[Region] ) ) RETURN IF ( CONTAINSSTRING ( CONCATENATEX ( temp_table, [TEST] ), "FALSE" ), "YES", "NO" ) - 3 years ago
Thanks MFelix for answering my question. Your measure works!
I didn't get a chance to post but I asked this question to ChatGPT and it gave me a similar formula but it didn't work. I changed ChatGPT's code a little and that works as well.
Fix Measure = CALCULATE(IF ( COUNTROWS ( FILTER ( ADDCOLUMNS ( SUMMARIZE ( 'test table', 'test table'[Item], 'test table'[Region], "SoldOutERP", SUM ( 'test table'[Sold Out ERP] ), "SoldOutWMS", SUM ( 'test table'[Sold Out WMS] ) ), "Mismatch", [SoldOutERP] <> [SoldOutWMS] ), [Mismatch] = TRUE() ) ) > 0, "Yes", "No" ) , ALL('test table'[Region]) )Thanks again!
Hello b2wise ,
can you explain when the fix column should show yes and when no, the condition that determines.
If I answered your question, please mark my post as solution so it would appeare to others, Appreciate your Kudos 👍
Idrissshatila Sorry if I wasn't clear. Compare the 1's and 0's in the two columns "Sold Out ERP" and "Sold Out WMS" by each item and region. If there are any mismatches the entire item should be marked as fix.
For item "Apple" all match hence "No", for Orange NJ and Pear NY there are discrepancies so the entire item should be marked as "Yes".