Compare Large Data Sets

[lora]

Can someone point me in the right direction!

We are retiring a server. I am transfering code to the new server and modifying sources. I need to compare the results of the data created from/on the old server to the new server. I used the compare task in SAS for a smaller data set - 55 variables, 1 million records and it worked great.

The data I'm having issues with has 313 variables, but only 85000 records...the compare task fails on memory. Proc freq on some variables does show that I have issues. I need a return data set with all of the variables/records that do not match.

Thanks in advance to anyone who can offer advice!

 

[klaz2002]

You could run the compare on half the data, twice. Or, you can trim the columns so that the compare is using less memory. Or you can dump another chip into the computer and the compare should run.

 

[DanJR]

you could also try set difference...

proc sql;
  create table blar as
  select * from table_1
  except
  select * from table_2;
quit;

and vice versa

cheers,
dan.

 

[lora]

Thanks to both of you. The except did work.

I finally got proc compare to work on my larger dataset(s) by modifying code generated by using the task list in EG from a smaller working compare.