Today at work I needed to move and rename a ton of files, but I also needed to make sure that the Git history followed. When I tried to move the files manually, SourceTree thought I deleted the files and then created new files elsewhere. When I commit files like this, SourceTree is usually smart enough to catch on, but it wasn’t happening. No good.

It took me a while to figure this all out, so I figured I’d share.

In the root, I had nearly 1000 files with the following naming schema: L1T1-page01.cfm (standing for lesson 1, topic 1, page 1). All together the project has 8 lessons, each lesson has several topics, and each topic has several pages.

I needed to move these files from the root and into organized folders: L1T1-page01.cfm would live at lessons/01-lesson-name/01-topic-name/L1T1-page01.cfm.

So I had to git mv these files myself; a new-to-me command.

First, I moved batches of files belonging to the same topic to their respective topic folders by hitting this command:

git mv L1T1* lessons/01*/01\*

(Can I get a hell yeah for that asterisk?)

After I moved the files I realized I wanted to shorten the filenames from L1T1-page01.cfm to 01.cfm.

Within each topic folder, I hit the following command:

for f in \*.cfm; do git mv $f $(echo \$f | sed ‘s/L[0-9]T[0-9]-page//g’);done

Update 1/7/2014: I revisited this post, I’m so glad I had save these commands! I found that instead of going into individual folders, I can run the commands from the lessons/ folder and let it ride:

for f in _/_/\*.cfm; do git mv $f $(echo \$f | sed ‘s/L[0-9]T[0-9]-page//g’);done

It worked. Just by looking at the last command, I know there’s a smarter way to write it. Still, it worked and I’m kind of proud of myself.

As I was figuring this out, I imagined the scene from Hook where Peter begins to remember how to play pretend and the children are like “You’re doing it, Peter.”