Proc STDRATE error message, missing matched stratum

[psh23]

I am trying to get directly standardized rates using proc stdrate, by census tract and county. When I run my code I get the error message "ERROR: The stratum in the REFDATA= data set, agegrp= 16, does not have a matched
stratum in the DATA= data set" . I added the "notsorted" option becuase I thought the problem might be that census tract is not in ascending order, but it did not fix the problem. 

proc sort data=test; by county1 ctract; run;

proc stdrate data=test
refdata=USstandard
method=direct
stat=rate(mult=100000)
;
population event=count total=denom;
reference total=standard;
strata agegrp;
by county1 ctract notsorted;
run;

[Reeza]

Check if your test dataset is missing a value for age 16 for a census tract. You can add it in as 0. 

A quick amd dirty way is to use proc freq with the sparse option to add in the 0. 

[psh23]

Hi Reeza,

I have run proc means n nmiss, and there are no missing values in the dataset, I think it must have something to do with the sorting of the data?

Thank you!!

[psh23]

It runs when I do not include census tract.

[Reeza]

It's not missing as in ., it's missing as there's no record for age 16. 

Try something like the following. 

Does it return the same value?

Proc SQL;

select count(distinct censusTract)

from have;

select count(distinct CensusTract)

from have

where age=16;

quit;

[Reeza]

Actually I missed that the error message identifies the exact group with the issue. Review your log and the error message again and check if that value exists in your dataset.

[psh23]

There are certain census tract / county strata where there are no values for a certain agegroup, because it is such a small subgroup, but those stratum are not included in the data. It seems like it is getting stuck in that it wants a value for every single age group stratum within each census tract / count. It is impossible to include these strata in the dataset because the denominator is zero and it will not run if there are zeros.