I'm trying to replace a string in multiple files. But only the first file gets replaced. The code is:
perl -pi -e 's/(^<section_begin>.*\s+$)/\1<expand>\n/ if $.==1' *exp`
grep -iw "expand" *exp
a3exp:<expand>
If I use the same command for an individual file, it's working:
perl -pi -e 's/(^<section_begin>.*\s+$)/\1<expand>\n/ if $.==1' n2exp
grep -iw "expand" *exp
a3exp:<expand>
n2exp:<expand>
Can you please help me to resolve this issue?
According to perlvar:
$. is reset when the filehandle is closed, but not when an open filehandle is reopened without an intervening close().
Because "<>" never does an explicit close, line numbers increase across ARGV files (but see examples in eof).
(second emphasis added)
When you use -p, @ARGV is set to the list of files you passed on the command line. Since $. is not reset when you go to the next file, it will only equal 1 for the first file. You can see this by simply printing $.:
$ echo foo > foo
$ echo bar > bar
$ perl -pe 'print "$. "' *
1 bar
2 foo
(with -p lines are printed automatically, so in the above example, the content of each line is printed next to the value of $.)
If you close the special filehandle ARGV when you move to the next file as mpapec suggests, $. will behave as you expect:
$ perl -pe 'print "$. "; close ARGV if eof' *
1 bar
1 foo
See the documentation for eof for more details and a couple of good examples.
By closing ARGV filehandle, line counter $. gets reset.
http://perldoc.perl.org/perlvar.html#ARGV
The special filehandle that iterates over command-line filenames in @ARGV
So you need to explicitly close file when eof() is reached,
perl -pi -e'
s/(^<section_begin>.*\s+$)/\1<expand>\n/ if $.==1;
close ARGV if eof;
' *exp
The Unix shell has a lot of tools to workaround this problem. There's find with -exec:
find . -name \*exp -exec perl -pi -e '... if $.==1' {} \;
or an old-fashioned for loop. In bash that would look like
for file in *exp
do
perl -pi -e '... if $.==1' file
done
xargsfind -exec approach will invoke the Perl interpreter once for each file. It would be better to pipe the results of find to Perl, something like: find -print0 | perl -p0i -e ... (assuming your version of find has -print0)You can use eof or eof(ARG) to detect the end of each file and reset $. accordinly.
From the Documentation
In a while (<>) loop, eof or eof(ARGV) can be used to detect the end of each file, whereas eof() will detect the end of the very last file only. Examples:
# reset line numbering on each input file
while (<>) {
next if /^\s*#/; # skip comments
print "$.\t$_";
} continue {
close ARGV if eof; # Not eof()!
}
I think your if $.==1 is getting you. That's lines processed: you're telling it to just process the first line of the input.
