1 minute read

Batch renaming in bash.

Often when I’m working on a revision of of a paper I find myself needing to rerun all of the analyses in the paper with some substantial tweaks. For example, the referee may have asked to see the main analysis with a new FE, cluster, or control structure (often all three).

When I’m working through this initially I’ll do the following to ‘sandbox’ the analyses.

First make a subdirectory in the analysis folder:

mkdir alternate_controls

Then I’ll move into that subdirectory and create copies of the main analysis. I tend to give the main analyses (those that get tables in the draft) a common prefix. In this case the analysis all estimate the effect of “aaca” on various metrics so they all start with aacaAnd.

cd alternate_controls
cp ../prefix*

Now, I want to quickly rename all the analyses to indicate that these are not the main analyses. You could do this one at a time, but you can also do it quickly using bash. here is a quick renaming script:

for file in *.do
	do mv $file ${file%.do}_alt_cntrls.do
done

the first line loops over all the matches for “*.do” in the folder. The do on the next line passes the shell command that follows for each “file” that ends in .do. mv is the unix command for ‘move’ which is equivalent to rename, $file is the original name. $file%.do is the origional name up to .do and then _alt_cntrls.do is the new suffix.

Tags:

Categories:

Updated: