linux - Remove non ASCII chars from multiple files recursive -

i have tons of files include non ascii characters , on complete file system. looking batch solution run 1 problem:

find . -name "*.yml" -print0 | while read -d $'\0' file; tr -cd '\11\12\15\40-\176' < "$file" > "$file"; done 

this command should work fine, wouldn't there problem tr -cd doesn't work if use same input , output name. know alternative or how solve small issue?

got (no idea if looks nice or crappy):

find . -name "*.yml" -print0 | while read -d $'\0' file; cp "$file" "${file}.temp" && tr -cd '\11\12\15\40-\176' < "${file}.temp" > "$file" && rm "${file}.temp"; done 

depending on file names processing, may want add ifs= , -r read command, see greg's bashfaq 001 details.

regarding question, use temporary file suggested carl or use sponge command the moreutils package. either:

find . -name "*.yml" -print0 |  while read -r -d '' file;    tr -cd '\11\12\15\40-\176' < "$file" > "${file}.tmp" \     && mv "${file}.tmp" "$file" done 

or:

find . -name "*.yml" -print0 |  while read -r -d '' file;    tr -cd '\11\12\15\40-\176' < "$file" | sponge "$file" done