Not Quite a Blog

(This post is mostly for me, as I end up doing things like this infrequently and tend to forget how to do it.)

I recently found myself with a CSV file that looked like so:

First,Last,Affiliation
....

and wanted to get it into a form for emailing:

First Last (Affiliation)
...

Emacs' replace-regexp is my tool of choice, so I messed around a bit and constructed the following regexp (should be all one unbroken line in emacs regexp minibuffer):

\([[:alnum:] ]+\),
\([[:alnum:] ]+\),
\([[:alnum:] ,\.\"\&\/\)\(]+\)

and replaced it with:

\1 \2 (\3)

The regexp does the following:

• \(...\) is a regexp container, that you can reference in the replacement with \1 (and for successive containers, use \2, \3, etc.)
• [[:alnum:] ]+ says find any sequence of characters containing A-Z, a-z or 0-9 or whitespace.
• [[:alnum:] ,\.\"\&\/\)\(]+ does the same but also includes a number of non-letter characters that people seem to use in their affiliations when presented with free-text entry, specifically [,."/)(]. (The comma doesn't need to be escaped... and we want to include it as someone may have an affiliation like "Univ. of Ca., Berkeley" where there's a comma inside the quoted string that's not a CSV delimiter.)

So, this is a long-winded way of saying: grab the first chunk of stuff before a comma, remember it; grab the next chunk of stuff before a comma, remember it; grab the rest of the stuff to the end of the line; remember it. And the replacement says, put the first chunk down then a space then the second chunk and a space then an open parenthesis, then the final chunk and finally a close parenthesis.