Yes, I saw your Excel file, however I will not touch Office files from the net, they are a security risk. Post test data - in the form of a datastep - in the body of your post, e.g.;
data 1;
input a b;
datalines;
1 1
2 2
;
run;
Now looking at what you have posted below, you have two tables, only one of which variable name matches - I have no idea on other properties such as format or length. Now what do you want to do with those two variables - this is where showing an want table is useful. It seems your looking to merge all the 70 tables into one really wide table, not a recommended approach. Much like your Oracle database, it makes life easier for programming if you keep the normalised structure (data goes down rather than across). Take this one, if you normalise the data to:
EMP_CODE PARAM RESULT
001 EMP_NAME Abc
001 DATE_OF_JOIN 01JAN2014
...
Then your task becomes trivial as you can just set all of the datasets into one big dataset. Later on if you need the transposed data, then use proc transpose or array processing to transpose the data up.
Also note that you can get one complete normalised dataset from the Oracle Clinical (OC) master dataset, i.e. avoiding all the retrieval of individual data and processing. In OC thre is a Questions, and reposnses master table, merge those two and export and it contains the param/result setup from above.
... View more
