March 31 - April 2, 2025, in Las Vegas, Nevada. Use code MSCUST for a $150 discount! Early bird discount ends December 31.
Register NowBe one of the first to start using Fabric Databases. View on-demand sessions with database experts and the Microsoft product team to learn just how easy it is to get started. Watch now
Hi,
I have a hard-coded PQ text parameter P.
There's a SQL database Table A, filtered where a field = P (so, 100 rows becomes 36, for example). This is good so far.
Here's the problem:
I need to filter SQL database Table B, which does not have a field holding P. I need to filter Table B based on the filtered content of Table A (so, 271 rows becomes 95, for example, where those 95 in Table B have a FieldX equal to the equivalent FieldX in Table A)).
I can create a Custom Step using the fx button . . . but what is the M code for this next filter step? I think it is going to be something like
#"Filtered Rows" = Table.SelectRows(TableB, each [FieldX] = ? )
in
#"Filtered Rows"
Am I on the right track?
All help gratefully received.
Solved! Go to Solution.
You really can do what @AlexisOlson advises, but if you'd rather use List.Contains, that's legit as well, like if you don't want the values from Table A (but my first thought was Inner Join as well):
Table.SelectRows(TableB, each List.Contains(List.Buffer(List.Distinct(TableA[FieldX])), [FieldX]))
--Nate
Thanks for this code.
I copied it very carefully, substituting the real names for the dummy ones, but failed dismally.
More by happenstance than skill, I arrived at
Table.SelectRows(dbo_TableB, each List.Contains(List.Buffer(List.Distinct(TableA[FieldX])), [FieldX]))
I can see what's happening, and why this is doing what I need.
dbo_TableB is referring to the underlying table in the database. That's good.
TableA is teferring to the filtered TableA which exists after the steps in the M code. That's also good.
Reading the M code as functional programming, and referring to the docs for each method, makes sense of it all. I've attached the real code for information.
let
Source = Sql.Database("redwinghealthserver.database.windows.net", "SMIPHC"),
dbo_PrimaryPatientObservation = Source{[Schema="dbo",Item="PrimaryPatientObservation"]}[Data],
JustThePatients = Table.SelectRows(dbo_PrimaryPatientObservation, each List.Contains(List.Buffer(List.Distinct(PrimaryPatient[NHSNumber])), [NHSNumber]))
in
JustThePatients
You can use that as the table to join with.
= Table.NestedJoin(
Source, {"FieldX"},
Table.SelectRows(TableA, each [P] = P), {"FieldX"},
"TableA", JoinKind.Inner
)
You really can do what @AlexisOlson advises, but if you'd rather use List.Contains, that's legit as well, like if you don't want the values from Table A (but my first thought was Inner Join as well):
Table.SelectRows(TableB, each List.Contains(List.Buffer(List.Distinct(TableA[FieldX])), [FieldX]))
--Nate
Thanks for this code.
I copied it very carefully, substituting the real names for the dummy ones, but failed dismally.
More by happenstance than skill, I arrived at
Table.SelectRows(dbo_TableB, each List.Contains(List.Buffer(List.Distinct(TableA[FieldX])), [FieldX]))
I can see what's happening, and why this is doing what I need.
dbo_TableB is referring to the underlying table in the database. That's good.
TableA is teferring to the filtered TableA which exists after the steps in the M code. That's also good.
Reading the M code as functional programming, and referring to the docs for each method, makes sense of it all. I've attached the real code for information.
let
Source = Sql.Database("redwinghealthserver.database.windows.net", "SMIPHC"),
dbo_PrimaryPatientObservation = Source{[Schema="dbo",Item="PrimaryPatientObservation"]}[Data],
JustThePatients = Table.SelectRows(dbo_PrimaryPatientObservation, each List.Contains(List.Buffer(List.Distinct(PrimaryPatient[NHSNumber])), [NHSNumber]))
in
JustThePatients
Thanks for this, and I appreciate what you're saying, but I can't do this in SQL, as the hard-coded parameter is a PQ parameter . . . all the filtering needs to be done on the client side.
Conceptually, what I want to do is JOIN the filtered TableA to TableB (or select from tableB using an EXISTS) and thus generate a filtered TableB based on the TableA subset.
The way forward looks like some sort of equivalent to an EXISTS like a List.Contains, but I dunno enough M to figure out what to do.
You can absolutely do an inner join in Power Query without any SQL involved. See the link I provided previously or take a look at this more general documentation about merges (joins):
https://docs.microsoft.com/en-us/power-query/merge-queries-overview
Ordinary JOINs are easy - no question.
But I have to admit defeat. I couldn't figure out how to do a JOIN in PQ on the filtered TableA. I couldn't figure out -- in the TableB query -- how to reference the data that existed in the TableA query after this step was applied:
= Table.SelectRows(dbo_TableA, each [P] = P) [where P is the PQ parameter]
You can use that as the table to join with.
= Table.NestedJoin(
Source, {"FieldX"},
Table.SelectRows(TableA, each [P] = P), {"FieldX"},
"TableA", JoinKind.Inner
)
This is closer to what I originally meant. It's also what a SQL person like myself would naturally gravitate to.
Thanks to everyone for a very useful discussion.
The way wrote it is similar to a subquery expression but you can also do it more like a CTE and define the filtered table as a prior step and then reference that step inside the join.