String Search/Replace

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

Replies are listed 'Best First'.
Re: String Search/Replace by Fletch (Bishop) on Aug 31, 2005 at 14:32 UTC
Fore! `perl -lane 'for(@F[1..4]){$_=0if$_>1000};print"@F"'` [download] Update: Shave a couple of strokes. `perl -lape 'for(@F[1..4]){$_=0if$_>1000};$_="@F"'` [download] -- We're looking for people in ATL	[reply] [d/l] [select]
Re^2: String Search/Replace by crep (Novice) on Aug 31, 2005 at 14:46 UTC
It worked perfectly! Thanks! -Jack C jack {@} crepinc.com http://www.crepinc.com/	[reply]
Re: String Search/Replace by ikegami (Patriarch) on Aug 31, 2005 at 14:24 UTC
Zillion? How about only one: `perl -ple "@f=split; for(@f[1..4]){ $_=0 if $_>1000 } $_=qq{@f}" infil +e >outfile` [download] or in-place: `perl -i.bak -ple "@f=split; for(@f[1..4]){ $_=0 if $_>1000 } $_=qq{@f} +" file` [download] Update: Better yet, `-a` saves us the trouble of splitting: `perl -lape "for(@F[1..4]){ $_=0 if $_>1000 } $_=qq{@F}" infile >outfil +e` [download] or in-place: `perl -i.bak -lape "for(@F[1..4]){ $_=0 if $_>1000 } $_=qq{@F}" file` [download] perlrun documents `-i`, `-l`, `-a`, `-p` and `-e`. The command can be written in fewer strokes, but I left it readable.	[reply] [d/l] [select]
Re^2: String Search/Replace by crep (Novice) on Aug 31, 2005 at 14:40 UTC
Hm... I tried the first in your original post and your updated, both give the same error. I check, and the file I'm using is definatly writable. `bash-2.05$ perl -ple "@f=split; for (@f[1..4]) { $_=1000 if $_>1000; } + $_=join(' ',@f)" TEST Can't modify constant item in scalar assignment at -e line 1, near "10 +00 if" Execution of -e aborted due to compilation errors. bash-2.05$ perl -lape "for(@F[1..4]){ $_=1000 if $_>1000 } $_=qq{@F}" +TEST Can't modify constant item in scalar assignment at -e line 1, near "10 +00 if" Execution of -e aborted due to compilation errors.` [download] Did I do something wrong? -Jack C jack {@} crepinc.com http://www.crepinc.com/	[reply] [d/l]
Re^3: String Search/Replace by ikegami (Patriarch) on Aug 31, 2005 at 14:41 UTC
I'm using quoting appropriate for the Windows shell. Substitute `"` for `'` and vice-versa for unix shells.	[reply] [d/l] [select]
Re^4: String Search/Replace by crep (Novice) on Aug 31, 2005 at 14:47 UTC
Re: String Search/Replace by gargle (Chaplain) on Aug 31, 2005 at 14:38 UTC
Hi, Many answers already, but, if you're looking to split up the numbers in variables: #!/usr/bin/perl use strict; use warnings; while (<DATA>) { chomp; my ($a, $b, $c, $d, $e) = split/ /; print "$a $b $c $d $e\n"; } __DATA__ 1125356700 0 0 0 0 1125356400 15 0 25 0 1125356100 25 0 25 0 1125355800 25 0 25 0 1125355500 25 0 25 0 1125397800 0 0 0 0 1125397500 0 0 0 0 1125397200 0 0 0 0 1125396900 0 0 0 0 1125396600 0 0 0 0 1125396300 0 0 0 0 1125396000 39006 0 63597 0 1125395700 63597 0 63597 0 1125395400 63597 0 63597 0 1125395100 63597 0 63597 0 1125394800 63597 0 63597 0 1125394500 63597 0 63597 0 1125394200 63597 0 63597 0 [download] -- if ( 1 ) { $postman->ring() for (1..2); }	[reply] [d/l]
Re: String Search/Replace by philcrow (Priest) on Aug 31, 2005 at 14:40 UTC
I like golf about as much as the next person, but how about something a bit more expansive: `#!/usr/bin/perl use strict; while (<>) { my ($timestamp, $number, @rest) = split; $number = 0 if ($number > 1000); print "$timestamp $number @rest\n"; }` [download] The solutions above work in place. This one is a filter. To use it do something like this: `./filter < input > output` [download] But don't make output the same as input or the shell will kill your data. Phil	[reply] [d/l] [select]
Re: String Search/Replace by ChrisR (Hermit) on Aug 31, 2005 at 15:07 UTC
How about a regex: #!c:\perl\bin\perl.exe -w use strict; use warnings; for my $line (<DATA>) { $line =~ s/ \d{4,}( \|$)/ 0$1/g; print $line; } __DATA__ 1125356700 0 0 0 0 1125356400 15 0 25 0 1125356100 25 0 25 0 1125355800 25 0 25 0 1125355500 25 0 25 0 1125397800 0 0 0 0 1125397500 0 0 0 0 1125397200 0 0 0 0 1125396900 0 0 0 0 1125396600 0 0 0 0 1125396300 0 0 0 0 1125396000 39006 0 63597 0 1125395700 63597 0 63597 0 1125395400 63597 0 63597 0 1125395100 63597 0 63597 0 1125394800 63597 0 63597 0 1125394500 63597 0 63597 0 1125394200 63597 0 63597 0 [download] Or for a one liner: `perl.exe -ple "s/ \d{4,}( \|$)/ 0$1/g" filename` [download] *Update:* Oops... As cheseter has pointed out, the above will replace 1000 when it should not. Here's my regex that works: `$line =~ s/ (\d+)/$1>1000?" 0":" $1"/ge;` [download] or at least I think it does.	[reply] [d/l] [select]
Re^2: String Search/Replace by ikegami (Patriarch) on Aug 31, 2005 at 15:33 UTC
`{4,}` is greedy, so the `( \|$)` is not needed. The following is sufficient, shorter and faster: `perl -ple "s/ \d{4,}/ 0/g" infile >outfile` [download]	[reply] [d/l] [select]
Re^3: String Search/Replace by ChrisR (Hermit) on Aug 31, 2005 at 15:49 UTC
Thanks. I didn't think about that. I've never been very good at golf (perl or otherwise).	[reply]
Re^3: String Search/Replace by Roy Johnson (Monsignor) on Aug 31, 2005 at 16:10 UTC
The `-l` is unnecessary, too. Caution: Contents may have been coded under pressure.	[reply] [d/l]
Re^2: String Search/Replace by chester (Hermit) on Aug 31, 2005 at 16:03 UTC
Actually doesn't the original question say to replace numbers "over 1000" with 0? Your solution also replaces 1000 with 0, which would be incorrect.	[reply]
Re^3: String Search/Replace by ChrisR (Hermit) on Aug 31, 2005 at 17:50 UTC
Err... You've got me there. I think you should remove the `strike` from your previous post. Your solution is the shortest since it appears that mine is incorrect.	[reply] [d/l]
Re^2: String Search/Replace by gargle (Chaplain) on Aug 31, 2005 at 18:38 UTC
Hi, I really do love regex's but sometimes, when only used to cut up a string in pieces, pieces seperated by the same character, I prefer to use split. Reading the code becomes a bit easier on the eyes because you can immediately deduct the meaning of the line of code. With regex's you still have to do a bit of mental juggling before the intention becomes clear. -- if ( 1 ) { $postman->ring() for (1..2); }	[reply]
Re: String Search/Replace by chester (Hermit) on Aug 31, 2005 at 14:57 UTC
Without using `split` (I think this is the shortest so far too...): `perl -ple "s/(?<!^)\b(\d+)/$1>1000?0:$1/eg" filename` [download] edit: Nope, see below. :) edit2: Never mind...	[reply] [d/l] [select]
Re: String Search/Replace by sh1tn (Priest) on Aug 31, 2005 at 14:49 UTC
Another one liner: `perl -lape 's/\S+\s+//&&s/(\d+)/$1>1000?0:$1/ge' filename` [download] Update: ok this (even shorter) does not remove the timestamp: `perl -pe 's/(?<=\s)(\d+)/$1>1000?0:$1/ge' filename` [download]	[reply] [d/l] [select]
Re^2: String Search/Replace by ikegami (Patriarch) on Aug 31, 2005 at 14:58 UTC
That removes the timestamp.	[reply]