Power BI is turning 10! Tune in for a special live episode on July 24 with behind-the-scenes stories, product evolution highlights, and a sneak peek at what’s in store for the future.
Save the dateEnhance your career with this limited time 50% discount on Fabric and Power BI exams. Ends August 31st. Request your voucher.
I am trying to use the method outlined in this video from BI Elite to slice on a measure value.
https://www.youtube.com/watch?v=AHiCE1N0XHE
As you can see, the slicer value provided is "80-90" and the value returned from my measure "Ranking_Views_Banding" is "80-90%".
The second column "Filter - View Bands" is a DAX measure:
Use Performance Analyzer to get the DAX code for the table visual and paste it into DAX Studio. There will be a variable with ValueFilter in the name, where it will likely be wrapping a SUMMARIZECOLUMNS inside a FILTER checking that the value of your filter view brands measure is Yes. Take that inner SUMMARIZECOLUMNS and evaluate it, that should give a better understanding of what is going on.
Thank you. This is the DAX query (with a couple of fields changed to stay anonymous!). I notice that if I evaluate __ValueFilterDM1 then I get the result I'm expecting - i.e. filtered for "Yes" on the "Filter - View Bands" measure.
How can I make PBI apply the same filter to the __DS0PrimaryWindowed variable?
// DAX Query
DEFINE
VAR __DS0FilterTable =
TREATAS({"YTD"}, 'Period A'[Is_YTD])
VAR __DS0FilterTable2 =
TREATAS({"LW-1"}, 'Period B'[Is_LW-1])
VAR __DS0FilterTable3 =
TREATAS({"Filter1"}, 'dim Product'[myAttributeFilter])
VAR __DS0FilterTable4 =
TREATAS({"90-100%"}, 'Slice by - View Bands'[View Band])
VAR __ValueFilterDM1 =
FILTER(
KEEPFILTERS(
SUMMARIZECOLUMNS(
'dim Product'[Item ID],
'Slice by - View Bands'[View Band],
'dim Product'[myAttribute1],
__DS0FilterTable,
__DS0FilterTable2,
__DS0FilterTable3,
__DS0FilterTable4,
"Ranking_Views_Banding", 'Web Detail'[Ranking_Views_Banding],
"v_Ranking_Views_Banding_FormatString", IGNORE('Web Detail'[_Ranking_Views_Banding FormatString]),
"Filter___View_Bands", 'Web Detail'[Filter - View Bands],
"v_Filter___View_Bands_FormatString", IGNORE('Web Detail'[_Filter - View Bands FormatString]),
"Ranking_Views_Percentile", 'Web Detail'[Ranking_Views_Percentile],
"v_Ranking_Views_Percentile_FormatString", IGNORE('Web Detail'[_Ranking_Views_Percentile FormatString]),
"v__Product_Views", 'Web Detail'[# Product Views],
"v___Product_Views_FormatString", IGNORE('Web Detail'[_# Product Views FormatString])
)
),
[Filter___View_Bands] = "Yes"
)
VAR __DS0Core =
SUMMARIZECOLUMNS(
ROLLUPADDISSUBTOTAL(
ROLLUPGROUP(
'dim Product'[Item ID],
'Slice by - View Bands'[View Band],
'dim Product'[myAttribute1]
), "IsGrandTotalRowTotal"
),
__DS0FilterTable,
__DS0FilterTable2,
__DS0FilterTable3,
__DS0FilterTable4,
__ValueFilterDM1,
"v__Product_Views", 'Web Detail'[# Product Views],
"Ranking_Views_Banding", 'Web Detail'[Ranking_Views_Banding],
"v_Ranking_Views_Banding_FormatString", IGNORE('Web Detail'[_Ranking_Views_Banding FormatString]),
"Filter___View_Bands", 'Web Detail'[Filter - View Bands],
"v_Filter___View_Bands_FormatString", IGNORE('Web Detail'[_Filter - View Bands FormatString]),
"Ranking_Views_Percentile", 'Web Detail'[Ranking_Views_Percentile],
"v_Ranking_Views_Percentile_FormatString", IGNORE('Web Detail'[_Ranking_Views_Percentile FormatString]),
"v___Product_Views_FormatString", IGNORE('Web Detail'[_# Product Views FormatString])
)
VAR __DS0PrimaryWindowed =
TOPN(
502,
__DS0Core,
[IsGrandTotalRowTotal],
0,
[v__Product_Views],
0,
'dim Product'[Item ID],
1,
'Slice by - View Bands'[View Band],
1,
'dim Product'[myAttribute1],
1
)
EVALUATE
__DS0PrimaryWindowed
//__ValueFilterDM1
ORDER BY
//[IsGrandTotalRowTotal] DESC,
[v__Product_Views] DESC,
'dim Product'[Item ID],
'Slice by - View Bands'[View Band],
'dim Product'[myAttribute1]
Bit of a shot in the dark but can you try commenting out __DS0Filter3 from both __ValueFilterDM1 and __DS0Core? I just wonder if the filter on the product table is interfering somehow.
Check out the July 2025 Power BI update to learn about new features.
User | Count |
---|---|
22 | |
7 | |
6 | |
6 | |
6 |
User | Count |
---|---|
27 | |
10 | |
10 | |
9 | |
6 |