Using `find` to exclude multiple paths matching patterns

Question

I'm trying to locate nested YAML and HTML files to replace a string with sed after creating the list; I'm not sure I fully understand how to use the -prune option, here's what I have:

find . -type f -name '*.yaml' -or -name '*.html' \
    -and -path './.git/*' -or -path '*/node_modules/*' -prune

For which I still get HTML files under node_modules directories

Answer 1

If you'd like to prune any directory called either .git or node_modules from the search tree, you would use

-type d \( -name .git -o -name node_modules \) -prune

This would cause find to not even enter these directories (the -type d is not strictly necessary, but I'll use it here for symmetry with -type f; see below).

Then you would add the other conditions,

-type d \( -name .git -o -name node_modules \) -prune -o \
-type f \( -name '*.yaml' -o -name '*.html' \) -print

Ending up with

find . \
    -type d \( -name .git -o -name node_modules \) -prune -o \
    -type f \( -name '*.yaml' -o -name '*.html' \) -print

Any action that you'd like to take on the pathnames that passes all tests should be done in place of the -print.

Note that the default logical operation between two predicates is -a (AND).

Answer 2

man find

Is a great place to look for more info on find.

This should work for you.

find . \
    -type f \
    '(' -name '*.yaml' -or -name '*.html' ')' -and \
    -not '(' -path './.git/*' -or -path '*/node_modules/*' ')'

Note the use of parentheses, and the need to quote them due to their being intrinsic built-in syntax to bash (and most other shells).