Shell Programming and Scripting

Note that if $14 sometimes contains more than 4 characters, the RE you're using may give you false positive matches (such as 91001, 19100, etc.) in addition to your desired match on 9100. Also, if any of your filenames could contain whitespace characters, piping through xargs could be dangerous. And, finally, since you're feeding the output from awk through wc -land your script is printing $0 and a count of the lines it prints, your final count will be the number of lines matched plus the number of times awk is invoked; not just a count of matching lines. If any of these are a concern, you could try the following script to just get the number of matching lines:

Code:

find . -name "*_201304*.csv" -type f -exec awk -F, '$14~/^(910[0-6]|915[0-3]|920[0-3]|930[0-5]|940[01]|951[01])$/{++cpt}END{print cpt+0}' {} +|awk '{cpt+=$1}END{print cpt+0}'

As always, if you want to try this on a Solaris/SunOS system, use /usr/xpg4/bin/awk, /usr/xpg6/bin/awk, or nawk instead of awk.

you can first get .csv list of files and then try to check on each file.
Before putting *.csv in the above code check it works for a single file.

the command brings back 1 as an output which makes in incorrect. i am expecting more than 3000 as as output


bash-3.2$ find . -name "*_201304*.csv" -type f -exec awk -F, '$14~/^(91[00-01]|92[00-01]|9302|9300])$/{++cpt}END{print cpt+0}' {} +|awk '{cpt+=$1}END{print cpt+0}'|wc -l
1

This is the third time in this thread that you've reported failure and each time it's been due to some modification that you've made. At least try the suggestion as provided before reporting that it fails. Then, if you can't get your modification to work, you can ask about it specifically.

There is no wc in Don's pipeline. His code prints only a single line with a count. So, naturally, wc will always return 1.

Regards,
Alister