Your SQL looks syntactically correct but if it also will return your desired result depends on your source tables. We can't tell if you don't share representative sample data and show us the desired result.  My first question would be: In your table tdcx.vCXComps how many rows (dates) do you have per PtnId?   As for keyword JOIN used on its own: Yes, this is the same like INNER JOIN. I'd consider it a good practice to always use the INNER for clarity.   If you use Proc SQL option FEEDBACK then the SAS log will show you to what the SQL compiler transforms your code for execution. In your case that's an implicit join using the WHERE clause only and with only AND conditions I assume it will return what you're after - but again: We can't really know without sample data and desired result.  28 proc sql feedback; 29 select sum(t1.theowin)/30 from work.slots t1 30 join tdcx.vcxcomps t2 31 on t1.ptnid=t2.ptnid 32 where datepart(t2.compdate) between '01jan2024'd and '31mar2024'd 33 and datepart(t1.datetimein) between datepart(t2.compdate) and intnx('month', datepart(t2.compdate), 1,'b'); NOTE: Statement transforms to: select SUM(T1.theowin) / 30 from WORK.SLOTS T1, TDCX.VCXCOMPS T2 where (T1.ptnid = T2.ptnid) and (DATEPART(T2.compdate) between '01JAN2024'D and '31MAR2024'D) and DATEPART(T1.datetimein) between DATEPART(T2.compdate) and INTNX('month', DATEPART(T2.compdate), 1, 'b'); 34 quit;   And just for completeness here the "proof" that SAS will convert JOIN to INNER JOIN 28 proc sql feedback; 29 select l.name 30 from sashelp.class l join sashelp.class r 31 on l.name=r.name 32 ; NOTE: Statement transforms to: select L.Name from SASHELP.CLASS L inner join SASHELP.CLASS R on L.Name = R.Name; 33 quit; ... View more

If TDCX points to an external database then it's quite possible INTNX won't get translated correctly into that database's SQL dialect. Id suggest trying SQL Passthru if you are reading an external database. ... View more

It's a bit hard to say where the error message comes from, but I think it is because you have a comma before FROM.   As for your "subselect" statements, they do not make sense at all. What are you trying to accomplish with those? ... View more

The method you show here may work with your input data, but don't count on it. PROC SQL does not guarantee that records are read in the same order, meaning that it is not certain that you will get the same output every time, even if your input data stays the same (SQL may work differently, and fetch data in a different order, if your system is changed in some way).   If your data is the same every month (and in the same order), a data step may guarantee that you get the same extract with the same seed.   In that case, you can do something like this the first month: data TejonOwned; set TDQT.vCXPlayer_SAS t1; WHERE t1.SubFightZone = "Tejon Owned"; if RANUNI(4321) <0.1 ; run; And in the second month, just move on to the next decile, with the same seed: data TejonOwned; set TDQT.vCXPlayer_SAS t1; WHERE t1.SubFightZone = "Tejon Owned"; if 0.1<=RANUNI(4321) <0.2 ; run; ... View more