I saw an obvious way to clean up my hacker_rank directory, and jumped into it. I had a series of directories from the Hacker Rank Python exercises in my root directory called collections_OrderedDict, collections_defaultdict, etc. And another series called itertools_groupby, itertools_combinations, etc.

I hadn’t specified which language the challenges were in, but thought this would fit in with the clean up. In the root hacker_rank directory, I made python_collections, python_itertools, and python_functionals directories, mirroring Hacker Rank’s categories.

I ran…

mv collections_* python_collections

…(and the same for itertools_*), which cleaned up my root directory a lot. Another abstraction layer would be to make a python directory, but right now I actually only do the Python exercises.

The Problem

Inside hacker_rank/python_collections/ I now had a series of directories named collections_..., which had become redundant.

I messaged Leo, and asked him to help me come up with a bash script to rename all of these in one swoop.

The Solution

This is what he came up with:

find . -maxdepth 1 -name 'collections*' -exec sh -c 'mv {} `echo {} | cut -d _ -f2`' \;

• find all files on this directory (.)
• recursing until a max depth of 1
• whose name matches 'collections*'
• on them exec the command sh -c
• which calls sh on the string passed as argument: ‘mv {} echo {} | cut -d _ -f2’
• which in turn says “call mv on the matched file ({}) and change its name to whatever is inside the backticks
• which is the matched name (echo {})
• which is then cut by the delimiter (-d) _
• and from that cut, you only return the second field: -f2

Leo says: “The ending \; is a find thing, I think it needs it every time you use -exec”