skipping lines in a file until regexp matches

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

Monks,

I'm having a problem using "until" while processing lines in a file. Specifically, I'm opening a file, and I want to skip all of the lines in it until I match a specific pattern; I then want to examine the rest of the file. Though I think this code should work, it doesn't (nothing is printed) ...

while (<LOG>) {

   do    { next }    until m/SCHEDULEREC OBJECT BEGIN SERVER_DAILY $da
+te/;
   print ("$. : $_") if    m/SCHEDULEREC OBJECT BEGIN SERVER_DAILY $da
+te/;
   print ("$. : $_") if    m/SCHEDULEREC OBJECT END SERVER_DAILY $date
+/;
   print ("$. : $_") if    m/ANS1228E Sending of object '(.*)' failed/
+ 

};
[download]

I know that the pattern in my "do ... until" statement exists, because if I comment that line out, the next line prints the match. I'm sure I'm misunderstanding the way "next" and "until" works. Can someone give me a clue, please?

Thanks ... BlueFlashlight

Comment on skipping lines in a file until regexp matches Download Code

Replies are listed 'Best First'.
Re: skipping lines in a file until regexp matches by Aristotle (Chancellor) on Dec 17, 2002 at 20:16 UTC
`do { BLOCK } until $cond;` executes the BLOCK at least once. Therefor, your `next` gets executed immediately and aborts the current iteration of the `while` loop. As a result, nothing ever gets printed. Further, the `until` indicates looping, but you don't iterate over anything in your BLOCK so it should be `if`. What you really want is a first loop that skips all the junk, and then a second one that does the work. `do { $_ = <LOG> } until m/SCHEDULEREC OBJECT BEGIN SERVER_DAILY $date/ or not define +d; do { print ("$. : $_") if m/SCHEDULEREC OBJECT BEGIN SERVER_DAILY $date/ +; print ("$. : $_") if m/SCHEDULEREC OBJECT END SERVER_DAILY $date/; print ("$. : $_") if m/ANS1228E Sending of object '(.)' failed/ $_ = <LOG>; } until not defined;` [download] Makeshifts last the longest.*	[reply] [d/l]
Re: skipping lines in a file until regexp matches by dpuu (Chaplain) on Dec 17, 2002 at 21:15 UTC
You could also try the bistable operator (".." in scalar context): `while (<LOG>) { next unless /START/..0; print; }` [download] will print all lines from the first that matches /START/. Use "`/START/../END/`" if you want to stop processing at the end of each section. --Dave	[reply] [d/l] [select]
Re: Re: skipping lines in a file until regexp matches by blueflashlight (Pilgrim) on Dec 18, 2002 at 04:41 UTC
Awesome! Thanks to everyone who responded; I went with dpuu's solution, as it seemed the easiest to "grok", and it worked perfectly! --blueflashlight	[reply]
Re: skipping lines in a file until regexp matches by jdporter (Paladin) on Dec 17, 2002 at 20:20 UTC
The following is not necessarily the way I'd write it from scratch, but as a first revision from your code, I'd suggest this: `while (<LOG>) { last if /SCHEDULEREC OBJECT BEGIN SERVER_DAILY $date/; } do { print "$. : $_" if /SCHEDULEREC OBJECT BEGIN SERVER_DAILY $dat +e/; print "$. : $_" if /SCHEDULEREC OBJECT END SERVER_DAILY $date/ +; print "$. : $_" if /ANS1228E Sending of object '(.)' failed/; } while <LOG>;` [download] jdporter ...porque es dificil estar guapo y blanco.*	[reply] [d/l]
Re: skipping lines in a file until regexp matches by runrig (Abbot) on Dec 17, 2002 at 20:27 UTC
Yet another way: `my $start_pattern = qr/.... $date/; my $end_pattern = qr/.... $date/; my $send_pattern = qr/.... whatever/; while (<LOG>) { if (my $status = /$start_pattern/..1) { print "$. : $_" if $status == 1 or /$end_pattern/ or /$send_pattern/; } }` [download] Update: That should probably be (the first version only prints the start pattern once, this version will print it every time it occurs): `while (<LOG>) { if (/$start_pattern/..1) { print "$. : $_" if /$start_pattern/ or /$end_pattern/ or /$send_pattern/; } }` [download] Another update: After dropping the status, I'd go with an answer with a 'next unless' like dpuu's answer below.	[reply] [d/l] [select]