Get certified for free when you join Fabric Data Days 2026 and dive into Fabric, Power BI, SQL, AI, and other essential data skills.
Join nowData Days is here! Join us now for 60+ days of learning, challenges, and connection. Learn more
To work on some tables, I found it useful to use the Table.FuzzyGroup function. Unfortunately the result of the tests carried out is different from what I would have expected.
In particular, the TransformationTable parameter seems to have a behavior that depends on the order of the rows of the table on which to make groupings.
It also seems to depend on the value assigned to the other parameter: Threshold.
Has anyone used this function with this parameter?
What do you think about the way it works?
Source 1
res 1
sorce 2
res 2:
source 3:
res 3:
Hi, @Anonymous
Power Query functions that contain the word Fuzzy are not easy to use, their Threshold parameter is particularly difficult to adjust, and we often don't get the results we expect using such functions.
Hi @ziying35 ,
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Column1", type text}, {"Column2", Int64.Type}}),
cols=Table.ColumnNames(#"Changed Type"),
n=List.Count(cols),
head=Table.InsertRows(#"Changed Type",0,List.Transform(List.Distinct(mapTab[To]), each Record.FromList(List.Repeat({_},n),cols))),
#"Grouped Rows" = Table.FuzzyGroup(head, {"Column1"}, {{"Count", each Table.RowCount(_)-1, Int64.Type}},[TransformationTable=mapTab])
in
#"Grouped Rows"
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Column1", type text}, {"Column2", Int64.Type}}),
cols=Table.ColumnNames(#"Changed Type"),
n=List.Count(cols),
head=Table.InsertRows(#"Changed Type",0,List.Transform(List.Distinct(mapTab[To]), each Record.TransformFields(Record.FromList(List.Repeat({null},n),cols),{cols{0},(x)=>_}))),
#"Grouped Rows" = Table.FuzzyGroup(head, cols{0}, {{"Count", each Table.Skip(_), Int64.Type}},[TransformationTable=mapTab])
in
#"Grouped Rows"
@Anonymous Have used it, has some limitations around just how much control you have over it. Yes, the order matters I believe as well as the threshold as I think the way the algorithm works is that it finds the first match that falls within the threshold essentially. Hard to know for certain as it is kind of a black box.
If you want more control, I did create this DAX Fuzzy matching Quick Measure Gallery submission here:
https://community.powerbi.com/t5/Quick-Measures-Gallery/Fuzzy/m-p/1352914#M608
Don't miss out on Data Days, June 15 through August 7. Learn Fabric, Power BI, SQL, AI and more.
Check out the May 2026 Power BI update to learn about new features.
| User | Count |
|---|---|
| 4 | |
| 4 | |
| 3 | |
| 2 | |
| 1 |
| User | Count |
|---|---|
| 11 | |
| 11 | |
| 5 | |
| 4 | |
| 4 |