interrogation signs in variable labels

[Melanie Marten]

Hello,

I have a dataset in french, and all "é" and "à" letters are interrogation signs in the variable labels.
The original dataset was a SAS dataset, and I exported the data directly as a Stata dataset.
Is there a way to automatically fix the interrogation signs in Stata or is this something to be done in SAS?

Many thanks!

[Rich Goldstein]

what do you mean "fix the interrogation signs"?

[Nick Cox]

If this means that some accents end up as question marks namely

? 

then the problem is that you can't easily reverse that. It's hard to say what's best. Assuming that you know enough French to know the right answers, then with 20 or so variables, it could be quickest just to get into varm and make the fixes. 200 or 2000 or ;;; variabes, not so. You might have to revisit the SAS export somehow.

[Melanie Marten]

Ok many thanks for the response! I do know the french and can easily figure out what each variable label says, I just find the interrogation signs annoying.
By "fix the interrogation signs" I meant revert back to the original letters, namely "é" or "à". My apologies for not being clear.

[daniel klein]

This is just a guess, but look into

help unicode

[Melanie Marten]

Thanks again!

I ended up just copying the summary of the dataset (after typing "describe"), and then used Notepad ++ to replace all the interrogation signs/ add the code to re-label the variables. It wasn't automatic, but it didn't take long!