12-05-2017 03:58 PM
I am trying to merge two files into a new combined file. Both files have many of the same variables. I am getting an error message on merging that says something to the effect "variables are both character and numeric." For example, there is one variable called "extract_month." In one file it is listed as "char" and in the other "numeric". (no idea how this happened!). I went into the attributes to change one from char to numeric but I can't change. I am only assuming this is the only problem with the merge. Any idea how to fix it? Thanks!
12-05-2017 04:02 PM
You can't change a variable type once it exists. You can drop the variable and recreate it however as a character/numeric.
However, it's usually easier to go back in your steps and fix that variable from the start.
Should the variable be numeric or character is the first question? - that's a business use/knowledge question that we can't answer
Why isn't it the correct type? - import from Excel where it assumed types, or from PROC IMPORT?
Where to fix it? - Ideally in the import step but you can also fix it after the fact.
If you do a SQL merge you can fix it in the join directly as well.
12-05-2017 04:06 PM
You have to convert one of the variables, so that both have the same type.
Also check the programs that created the two datasets, maybe you can already change it there.