UNIX for Dummies Questions & Answers

Hi ,
I am using the following script for finding and global replace

perl -p -i -e "s/gruesome/wonderful/g" testfile.lst

the above command is working if i execute on shell.......but it is not working when place it in a perl script and exeute it...i am getting a error "BASEWORD found where operator expected"

i used

perl -p -i -e "s/gruesome/wonderful/g" testfile.lst;

'perl -p -i -e "s/gruesome/wonderful/g" testfile.lst';

'perl -p -i -e s/gruesome/wonderful/g testfile.lst';

in the file but none is working......what am i doin wrong....i 've been struck with this too......

Regards
Vivek.S

If you read the perlrun manpage you will know how to convert your single-liner into a proper Perl script. A quick snippet:

Code:

rename "testfile.lst", "testfile.lst.tmp";
open FILE_IN, "<testfile.lst.tmp" || die "Cannot open file for reading";
open FILE_OUT, ">testfile.lst" || die "Cannot open file for writing";
while (<FILE_IN>) {
s/gruesome/wonderful/g;
print FILE_OUT $_;
}
close FILE_IN; close FILE_OUT;
unlink "testfile.lst.tmp";

Code:

$ chmod +x test.pl
$ more oldfile
This is gruesome
some gruesome stuff
how gruesome this is
$ more test.pl
#!/usr/bin/perl
open( INFILE, "< oldfile" );
open( OUTFILE, "> newfile" );
while( <INFILE> ) {
  $_ =~ s/gruesome/wonderful/g;
  print OUTFILE $_;
}
close( INFILE );
close( OUTFILE );
rename( "newfile", "oldfile" );
exit( 0 );
$ ./test.pl
$ more oldfile
This is wonderful
some wonderful stuff
how wonderful this is

EDIT> cbkihong looks like we were typing our replies at the same time

Cheers
ZB

THANKS A LOT GUYS!!
I never knew there was so much to be done!!

why does this alone does not run while grep and other nix commands run on the shell script?

is there any specific reason?

Thanks and Regards
Vivek.S

Yep. But I was lucky to hit the "submit" button sooner.


The simple answer:
Perl and awk/sed are different things.

The Perl command-line arguments offer some switches to make shell scripters happy. Because many shell scripters particularly on Unix expect to write a simple single-liner for various sysadmin-related file processing (as with awk/sed) instead of "typical" programming, the -p and -i arguments makes it convenient to do in-place file editing with Perl as you expect, say, in sed. But we all know this is not real in-place editing, so the rename-read-write strategy is actually used. These options make it possible for you to emulate the command-line behaviour of sed/awk. With the -e argument, it can be used by, say, Makefiles to construct the replacement pattern on-the-fly without preparing the script file in advance.

Perl is a more general programming language than awk/sed. Think of it as something like Java, C++ or PHP. They are not like awk/sed which are specific for file processing, so in-place edit are not made the default. And you need to spell it out explicitly if you are to do it in any programming languages, including Perl.

THANKS A LOT YAAR!

it helped to improve my understanding of programming in perl!!!

thanks and regards
vivek.s