process file via command line

fireartist has asked for the wisdom of the Perl Monks concerning the following question:

Hi,
I wish to process a file with a short command line script (*nix).
I don't have experience of running perl from the command line like this.
After reading perl --help I tried.
perl -i ./access_log -e 'while (<>) { #stuff# }' but from the error messages I could tell that perl was trying to execute the file 'access_log'.
I then tried
less access_log > perl -e 'while (<STDIN>) { #stuff# }' but from the output could tell that the file wasn't being read.

I've had a search around here, but can't find the answer.
Could someone shed some light on this?
Thanks

Comment on process file via command line Select or Download Code

Replies are listed 'Best First'.
Re: process file via command line by broquaint (Abbot) on Nov 11, 2002 at 17:31 UTC
Try this instead `perl -i -pe '# your code here' access_log` [download] See `perlrun` for more info about munging from the command line. HTH `_________ broquaint`	[reply] [d/l]
Re: process file via command line by robartes (Priest) on Nov 11, 2002 at 17:28 UTC
You want to do this: `cat access.log \| perl -ne '#stuff#'` [download] Or, if you want to print the result after every line: `cat access.log \| perl -nep '#stuff#'` [download] The `-n` switch puts the `while (<>)` loop around your input file for you: see perlrun. And, to really spoil you for choice, Perl is smart enough to decide where its input is coming from, so you can do this as well: `perl -ne '#stuff#' access.log` [download] CU Robartes-	[reply] [d/l] [select]
Re^2: process file via command line by Aristotle (Chancellor) on Nov 15, 2002 at 17:59 UTC
That is a useless use of `cat`.. :) Makeshifts last the longest.	[reply]
Re: process file via command line by Enlil (Parson) on Nov 11, 2002 at 17:28 UTC
you are probably looking for something like the -n or -p option. Like `perl -ni.bak -e 'stuff' input_file` [download] There is a good reference as to the use of these perlrun in the documentation. -enlil	[reply] [d/l]
Re: process file via command line by Bird (Pilgrim) on Nov 11, 2002 at 17:33 UTC
You'll probably be interested in the command line switches -p, -n, and -l... These give various levels of convenience when dealing with a file from the command line (see perlrun). You'll then want to use -e, as you've done, but follow the quoted code with the filename you're working on, so that it's fed to your script as a parameter. `perl -ne 'print unless /^#/' myfile.log` [download] -Bird	[reply] [d/l]
Re: process file via command line by fireartist (Chaplain) on Nov 12, 2002 at 08:39 UTC
Thank you robartes, Enlil, broquaint and Bird , you were all incredibly quick off the mark, and very helpful too! ++	[reply]