[SOLVED] Stuck on perl Regex problem need help?

Gary Baker · 08-24-2016, 09:26 AM

#!/usr/bin/perl

while($line = <STDIN>) {
$line =~ s/^.*$/"\^$&"/;
print ($line);
}

I have the above. What I am trying to do is to add quotes to the original output. The above adds quotes but also adds the caret. Thanks in advance.

keefaz · 08-24-2016, 12:02 PM

Code:

#!/usr/bin/perl

while(<>) {
  s/.*/"$&"/;
  print;
}

pan64 · 08-24-2016, 12:25 PM

why regexp?

Code:

$line="\"$line\"";

keefaz · 08-24-2016, 03:46 PM

Yes, could also use qq function to avoid escaped quotes

Code:

$line = qq/"$line"/;

sundialsvcs · 08-24-2016, 05:24 PM

Quote:

Originally Posted by Gary Baker

$line =~ s/^.*$/"\^$&"/;
}

I have the above. What I am trying to do is to add quotes to the original output. The above adds quotes but also adds the caret. Thanks in advance.

It appears to me that you are telling Perl that you specifically want(!) that caret character!

The first part of the s/from/to/ regex is ... well ... a pedantic way to say: "the entire string."

The second part then says to substitute:

A double-quote.
The character "^", which has been escaped with a backslash to be sure that Perl sees it as a literal character.
The Perl variable $&.
A double-quote.

"You asked for it, so you got it."

Rinndalir · 08-24-2016, 06:11 PM

Do like @pan64 posted. Larry himself does not recommend regex. But the other point of this is to go with the simplest approach. Regex is great and perl really cemented into the minds of many but it's not often simple.

sundialsvcs · 08-24-2016, 06:16 PM

"Meh ... who really cares what The Larry™ thinks?"

But, "yes, certainly ..." there was no actual reason to use regular expressions in this case.

Rinndalir · 08-25-2016, 12:39 PM

Quote:

Originally Posted by sundialsvcs

But, "yes, certainly ..." there was no actual reason to use regular expressions in this case.

May be there was, only the OP knows certainly. But maybe someone posts a benchmark to show the results for a million goes of each.

Gary Baker · 08-26-2016, 04:16 AM

Thanks guys I don't know why I use regex. The answers were great!

sundialsvcs · 08-26-2016, 10:00 AM

Quote:

Originally Posted by Rinndalir

May be there was, only the OP knows certainly. But maybe someone posts a benchmark to show the results for a million goes of each.

"A million goes of" anything at all is basically going to be an exercise in brutalizing the memory-manager.

String manipulation, no matter how you do it, can be very brutal because it "fills the air with sawdust and wood chips." I once worked with a high-performance section of an application which purposely used C(!) to do certain string operations ... in fixed arrays. It accepted dynamic-string inputs, dutifully tested them for excessive length, then performed a variety of operations in the "old school" way, purposely avoiding the dynamic allocation of memory. The working-set size (amount of VM typically allocated to and needed by the process) was dramatically reduced, and performance enhanced on the machine as a whole ... which ran multiple copies of this system at the same time.

grail · 08-26-2016, 10:52 AM

Please remember to mark as SOLVED once you have your solution.

jpollard · 08-26-2016, 11:08 AM

Quote:

Originally Posted by sundialsvcs

"A million goes of" anything at all is basically going to be an exercise in brutalizing the memory-manager.

String manipulation, no matter how you do it, can be very brutal because it "fills the air with sawdust and wood chips." I once worked with a high-performance section of an application which purposely used C(!) to do certain string operations ... in fixed arrays. It accepted dynamic-string inputs, dutifully tested them for excessive length, then performed a variety of operations in the "old school" way, purposely avoiding the dynamic allocation of memory. The working-set size (amount of VM typically allocated to and needed by the process) was dramatically reduced, and performance enhanced on the machine as a whole ... which ran multiple copies of this system at the same time.

:-)
That is the usual solution to C++ problems as well... as well as any language using dynamic strings.

Rinndalir · 08-26-2016, 03:24 PM

Code:

First Method:  3 wallclock secs ( 2.46 usr +  0.61 sys =  3.07 CPU) @ 291757.98/s (n=895697)
Second Method:  3 wallclock secs ( 0.79 usr +  2.47 sys =  3.26 CPU) @ 237464.11/s (n=774133)
Third Method:  3 wallclock secs ( 1.79 usr +  1.26 sys =  3.05 CPU) @ 73415.41/s (n=223917)
Underdog Method:  4 wallclock secs ( 2.68 usr +  0.42 sys =  3.10 CPU) @ 338334.84/s (n=1048838)

Will post the methods later.