Is there an option to add the N statistic onto PROC MIXED?

[Buzzy_Bee]

On this instructional document about PROC MIXED:

https://www.pharmasug.org/proceedings/2021/SA/PharmaSUG-2021-SA-062.pdf

They demonstrate the example below to create an output that shows Least Square Means and columns for Estimate, Standard Error, DF, Lower and Upper. I need all of those, but I also want the N statistic. Does anyone know if there is an option that can be added in so I can get N on this output please? Thanks.

ods output lsmeans=pb_lsmean diffs=pb_lsdiff;
 proc mixed data=qlqc2 method=reml covtest empirical;
by param;
class subjid trt visit;
model chg=base trt visit trt*visit;
random intercept/ subject=subjid;
repeated visit/ subject=subjid type=ar(1);
lsmeans trt*visit/ cl pdiff;
 run;

[data_null__]

Consider displaying  Univariate(PROC MEANS) and MIXED statistics in rows.  

[Buzzy_Bee]

I'm not sure what you mean by this. Industry convention requires statistics to be shown in tables where each statistic has a vertical column that displays the results and the treatments are listed horizontally in the table. If I displayed it vertically, no one would be able to read it or use in their reports.

[Rick_SAS]

Add the NObs table to your ODS OUTPUT statement:

ods output lsmeans=pb_lsmean diffs=pb_lsdiff nobs=pb_nobs;

[Buzzy_Bee]

That only shows me the number of observations in the data set though, which Proc Mixed already does by default. What I mean is, I need the N statistic. So in the LSM table where it shows Treatment Drug A and the Estimate, DF etc, I need an N column in there that would tell me there were say 100 for Drug A, and 120 for Drug B.
The Proc Means procedure shows N, so currently I'm creating a Proc Mixed followed by Proc Means to generate all of the statistics I need. But I wanted to create a nice, tidy table that shows them all together that I can easily copy and paste onto an MS Word document.

[SteveDenham · Accepted solution]

You are going to have to merge the PROC MEANS output with the PROC MIXED output.  PROC SQL is probably the best tool for that.

SteveDenham