Interpretation of Errors from PROC REG - No valid observations found

[Radwan]

Hello 

when I run this code 

proc reg noprint data=found1 outest=regress1;
model TACC1=  part1 part2 part3/ r;
by year induscode; run;

the sas program gave me this message 

NOTE: The above message was for the following BY group:
year=2008 indusCode=G60
ERROR: No valid observations are found.
NOTE: The above message was for the following BY group:
year=2009 indusCode=C24
ERROR: No valid observations are found.
NOTE: The above message was for the following BY group:
year=2016 indusCode=E49
ERROR: No valid observations are found.
NOTE: The above message was for the following BY group:
year=2017 indusCode=B10
ERROR: No valid observations are found.
NOTE: The above message was for the following BY group:
year=2017 indusCode=O80
ERROR: No valid observations are found.
NOTE: The above message was for the following BY group:
year=2017 indusCode=R88
NOTE: Interactivity disabled with BY processing.
NOTE: PROCEDURE REG used (Total process time):
real time 0.80 seconds
cpu time 0.68 seconds

NOTE: The data set WORK.REGRESS2 has 1296 observations and 11 variables.

 I got the results  but i m doubted about this message.

please what does it mean?

[PeterClemmensen]

This happens because you do not have valid data for some of your by groups. Below is a very simple example where I added an observation to the sashelp.class data set with the sex variable value 'T'. Then I set the height and weight variables to missing and run a similar regression and use sex as a BY variable..

data regdata;
   set sashelp.class end=eof;
   output;
   if eof then do;
      SEX='T';
      call missing(height, weight);
      output;
   end;
run;

proc sort data=regdata;
   by sex;
run;

proc reg data=regdata noprint;
   model weight=height;
   by sex;
run;

To clarify..  You will get regression results to interpret on. But not for the by-groups that do not have valid data.

[Radwan]

Ok but what kind of non validity of  the data?

[PeterClemmensen · Accepted solution]

Well, I can't see your data. But you have to dig in a little deeper to find out. For example, what data do you get when you run this?

data test;
   set found1;
   where year=2008 and indusCode='G60';
run;

If it is all missing for your regression variables, then you have your answer

[Reeza]

FYI - I've updated your subject here to make it more descriptive. Please do the same to your other open post: https://communities.sas.com/t5/SAS-Procedures/a-new-question/m-p/532382