Outputting Residual values with BY statement

[RandoDando]

I am running a regression model with a BY statement, where individual regression models are run for each group.  I want the residual values which are stored in the output dataset to be for the regression model of the applicable group for each observation.  I thought that just by placing the BY statement before the output statement would do this, but after looking at the residual values that doesn't appear to be the case.  Is there a way to do this within the PROC REG procedure or will separate output groups have to be concatenated?  If the latter is the case, how do I get an output table for each model?  Here is how I wrote my code:

proc reg data=MY_DATA ;
model  Y_VAR = X_VAR/r ;
by GRP_NAME;
output out=resid
r=yresid 
cookd=cooks;
run;

[PaigeMiller · Accepted solution]

Works for me. Show us (a portion of) the output data set named RESID so we can see what you are seeing.

[RandoDando]

@PaigeMiller  After I sent this question, I ran PROC REG on 1 group, then output that as RESID1.  I then merged that with the applicable subset of RESID by ID and compared the residual values.  They matched.  

A couple of the groups have higher than expected residuals for some cases, and that is what threw me off at first.  

[Reeza]

REG needs a QUIT at the end because it can be interactive.