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
I've been busting my head for a few days and cannot seem to accomplish what I feel should be a trivial task.
I have a table with one of the columns being a structured column (Table) looking like this:
Fig. 1
The same table has another basic text column containing:
Fig. 2
I'd like to create a new column with the Custom2 (Fig. 1) content filtered by the PerformanceIndicatorsParsed (Fig. 2).
Expected result for 1st row:
For 4th row:
Here is the code:
let
filter_indicator = () =>
let
output = [PerformanceIndicatorParsed]
in
output,
Filtered = Table.SelectRows([Custom2], filter_indicator()),
Result = Table.FirstN(Filtered,10)
in
Result
Unfortunately, i receive the below error:
Expression.Error: We cannot convert the value "each [PI_Id] = 2 or ..." to type Function.
Details:
Value=each [PI_Id] = 2 or [PI_Id] = 3 or [PI_Id] = 1
Type=[Type]
If i hardcode the condition like below, it works just fine.
let
filter_indicator = () =>
let
output = each [PI_Id] = 2 or [PI_Id] = 3
in
output,
Filtered = Table.SelectRows([Custom2], filter_indicator()),
Result = Table.FirstN(Filtered,10)
in
Result
Any help is greatly appreciated.
Regards
Solved! Go to Solution.
Hi @brassy27,
you were almost there. All you needed was Expression.Evaluate overlay:
let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUUpNTM5QiPZMibU1VMgvArMUbBWMlGJ1opWMkOVjSg0MjJON4IrA/FQFU7BKY6BKMMMMxjAHM2IB", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Id = _t, Filter = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Id", Int64.Type}, {"Filter", type text}}),
#"Added Custom" = Table.AddColumn(#"Changed Type", "Custom", each Table.SelectRows(#"Changed Type", Expression.Evaluate([Filter])))
in
#"Added Custom"
Cheers,
John
Hi @brassy27,
you were almost there. All you needed was Expression.Evaluate overlay:
let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUUpNTM5QiPZMibU1VMgvArMUbBWMlGJ1opWMkOVjSg0MjJON4IrA/FQFU7BKY6BKMMMMxjAHM2IB", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Id = _t, Filter = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Id", Int64.Type}, {"Filter", type text}}),
#"Added Custom" = Table.AddColumn(#"Changed Type", "Custom", each Table.SelectRows(#"Changed Type", Expression.Evaluate([Filter])))
in
#"Added Custom"
Cheers,
John
Oh yeah!!! Thanks a million John, you nailed it!
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 |