Is that possible to find files reside several level down in awk - awk

Question

May I know is there a way to find certain synctax of words in awk using one liner command whereby the files I'm finding is on deeper hier. Wondering if 'find" cmd can apply here but I tried with find it is not working for me, probably my usage is wrong.
My command as below and it works beautiful if I placed the command on the same level where the *.abc reside.
awk '/hdl_file/{printf "%s", $0}/input|output/{printf "%s", $0}' *.abc | sed s/\.aux// | sed s/,//g | sed s/input// | sed s/output// | sed s/=// | sed s/\"//g | sed s/\.ac// | sed s/hdl_file// | awk '{print $NF $0}' | awk '{$NF="";print $0}'
However, I need to apply this command to looks for *.abc file on 2 hierarchy deeper. Tried command below with find but it gave error.
find . -mindepth 2 | awk '/hdl_file/{printf "%s", $0}/input|output/{printf "%s", $0}' *.abc | sed s/\.aux// | sed s/,//g | sed s/input// | sed s/output// | sed s/=// | sed s/\"//g | sed s/\.ac// | sed s/hdl_file// | awk '{print $NF $0}' | awk '{$NF="";print $0}'
awk: No match.
Pls help. Thanks.

Answer 1

You could also use
awk '....' */*/*.abc
if the files are exactly 2 subdirectories deep. For arbitrary depth, a recent bash can do
shopt -s globstar
awk '....' **/*.abc
Your pipeline of awks and seds can be reduced to
awk '
/hdl_file/ || /input|output/ {
gsub(/[,"]/, "")
sub(/\.aux|\.ac|input|output|hdl_file|=/, "")
last = $NF
$NF = ""
print last, $0
}
' *.abc

Answer 2

I'm not sure if that's your intention, but -mindepth will make find search in directories that are more than 2 levels down. If you want files that are at most 2 levels down, you need to use -maxdepth.
If you want to use the output of one command (in this case find) as arguments for another command (awk), you can use xargs, for example:
find . -maxdepth 2 | xargs awk '[your awk script goes here]'
Or, you could use -exec to read the files, if you don't care about their names:
find . -maxdepth 2 -exec cat '{}' \; | awk '[your awk script goes here]'
# Or possibly:
find . -maxdepth 2 | xargs cat | awk '[your awk script goes here]'
Note also, that you can avoid spawning multiple instances of sed and piping them together, here's two ways to do that:
sed -e 's/foo/bar/' -e 's/something/somethingelse/'
sed -f sed_script # sed_script is a text file with commands for sed