A way to avoid repeated conditional loops

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

Replies are listed 'Best First'.
Re: A way to avoid repeated conditional loops by tinita (Parson) on Aug 24, 2011 at 11:18 UTC
use the flip-flop operator. if you want to read from alpha until end of file: `while (<$fh>) { if ( (m/^alpha$/ .. eof($fh) ) > 1) { # everything between alpha and end of file } }` [download] you can even change that second condition to 0: `if ( (m/^alpha$/ .. 0 ) > 1)`	[reply] [d/l] [select]
Re^2: A way to avoid repeated conditional loops by Deus Ex (Scribe) on Aug 24, 2011 at 12:44 UTC
That's an awesome approach! Really like this method. I'll test it and put it on production. Thanks!	[reply]
Re^3: A way to avoid repeated conditional loops by SuicideJunkie (Vicar) on Aug 24, 2011 at 13:57 UTC
Are you assuming that 'alpha' is also always the first line in the file? Also, the posted flipflop doesn't behave the same as the OP A simpler way to just use a flag and short circuit the regex work: `my $seen = 0; while (<$fh>) { if ( !$seen and m/^alpha$/) { $seen = 'yes indeed'; print "true\n"; }else{ print "false\n"; } }` [download] Given a test file of: `foo bar alpha beta gamma delta` [download] The results I got are: `Original: false false true false false false Flipflop: Argument "" isn't numeric in numeric gt (>) at test.pl line 21, <$fh> +line 1. false Argument "" isn't numeric in numeric gt (>) at test.pl line 21, <$fh> +line 2. false false true true true Flag: false false true false false false` [download]	[reply] [d/l] [select]
Re^4: A way to avoid repeated conditional loops by tinita (Parson) on Aug 24, 2011 at 15:10 UTC
Re^5: A way to avoid repeated conditional loops by Anonymous Monk on Aug 24, 2011 at 16:27 UTC
Re^4: A way to avoid repeated conditional loops by Deus Ex (Scribe) on Aug 24, 2011 at 14:43 UTC
Re: A way to avoid repeated conditional loops by AR (Friar) on Aug 24, 2011 at 10:37 UTC
`... while ( <FILE> ) { if ( /alpha/ ) { some code #1 last; } else { some code #2 } } while ( <FILE> ) { some code #2 } ....` [download]	[reply] [d/l]
Re^2: A way to avoid repeated conditional loops by JavaFan (Canon) on Aug 24, 2011 at 10:57 UTC
I wonder whether the copying of `some code #2` is worth the avoidance of the test. Another way of writing it, with the same duplication: `while (<FILE>) { if (/alpha/) { some code #1; while (<FILE>) { some code #2; } } else { some code #2; } }` [download]	[reply] [d/l] [select]
Re^2: A way to avoid repeated conditional loops by ikegami (Patriarch) on Aug 24, 2011 at 17:55 UTC
You can't rely on readline returning EOF or errors more than once. Fix: `OUTER: { for (;;) { defined( my $_ = <$fh> ) or last OUTER; if (/alpha/) { f(); last; } else { g(); } } while (<$fh>) { g(); } }` [download] Update: The remainder of this post is wrong. You can get rid of the duplicate g() by making the loops bottom tested. `OUTER: { defined( my $_ = <$fh> ) or last OUTER; for (;;) { if (/alpha/) { f(); last; } defined( $_ = <$fh> ) or last OUTER; } for (;;) { g(); defined( $_ = <$fh> ) or last OUTER; } }` [download] You can squish it down a bit: `LOOP: { defined( my $_ = <$fh> ) or last; if (/alpha/) { f(); redo; } for (;;) { g(); defined( $_ = <$fh> ) or last; } }` [download] Look at that, g() is not even duplicated anymore!!	[reply] [d/l] [select]
Re^3: A way to avoid repeated conditional loops by jffry (Hermit) on Aug 24, 2011 at 19:39 UTC
Why is your `for` loop idiom better than AR's or JavaFan's `while` loop statements above?	[reply]
Re^4: A way to avoid repeated conditional loops by ikegami (Patriarch) on Aug 24, 2011 at 20:36 UTC
Re^5: A way to avoid repeated conditional loops by jffry (Hermit) on Aug 24, 2011 at 21:46 UTC
Re^3: A way to avoid repeated conditional loops by Deus Ex (Scribe) on Aug 25, 2011 at 13:41 UTC
The squished version is really cool :) It looks I got my solution! Many thanks ikegami!	[reply]
Re^4: A way to avoid repeated conditional loops by ikegami (Patriarch) on Aug 25, 2011 at 17:45 UTC
Re^5: A way to avoid repeated conditional loops by Deus Ex (Scribe) on Aug 29, 2011 at 03:44 UTC
Re^2: A way to avoid repeated conditional loops by runrig (Abbot) on Aug 24, 2011 at 17:00 UTC
If 'some code #2' is very small or can be packaged into a single function call, I might go with that. Or I might just make it obvious that we only want/expect this once in the one loop: `my $got_alpha; while (<FILE>) { unless ($got_alpha) { if (/alpha/) { #...alpha code $got_alpha++; next; } } # ...more code }` [download] Update: And I just noticed the part in the OP where it says "without a flag variable"...oh, well...	[reply] [d/l]
Re^3: A way to avoid repeated conditional loops by AR (Friar) on Aug 24, 2011 at 17:04 UTC
I think you missed the point of the original post. Deus Ex wants to get rid of running the conditional as soon as "alpha" is seen. Your code substitutes that for another conditional (that doesn't use the regex engine, but still).	[reply]
Re^2: A way to avoid repeated conditional loops by Deus Ex (Scribe) on Aug 24, 2011 at 10:43 UTC
Interesting. I would never think of doing like this. Thanks :) The only point, at first glance, would be that you get twice the "while" loop. I'm trying to understand if you can avoid such "ugly" (take the word for it) structures. Thanks though.	[reply]
Re^2: A way to avoid repeated conditional loops by ikegami (Patriarch) on Aug 24, 2011 at 17:43 UTC
You can't rely on readline returning EOF or errors more than once.	[reply]
Re: A way to avoid repeated conditional loops by Eliya (Vicar) on Aug 24, 2011 at 19:37 UTC
Yet another way would be to set up a function(ref) which you just call for every line. Initially, the called function checks for /alpha/, but once alpha is encountered, it replaces itself with the routine do_beta that is then called directly for the remaining beta lines. That way, the condition is no longer tested from there on. `my $f; $f = sub { if (/alpha/) { do_alpha(); $f = \&do_beta; } else { do_beta(); } }; while (<DATA>) { $f->(); } sub do_alpha { print "do_alpha(): \$f = $f, \$_ = $_"; } sub do_beta { print "do_beta(): \$f = $f, \$_ = $_"; } __DATA__ beta beta beta beta alpha beta beta beta beta beta beta beta beta` [download] Output: do_beta(): $f = CODE(0x796e88), $_ = beta do_beta(): $f = CODE(0x796e88), $_ = beta do_beta(): $f = CODE(0x796e88), $_ = beta do_beta(): $f = CODE(0x796e88), $_ = beta do_alpha(): $f = CODE(0x796e88), $_ = alpha do_beta(): $f = CODE(0x7c6ee8), $_ = beta do_beta(): $f = CODE(0x7c6ee8), $_ = beta do_beta(): $f = CODE(0x7c6ee8), $_ = beta do_beta(): $f = CODE(0x7c6ee8), $_ = beta do_beta(): $f = CODE(0x7c6ee8), $_ = beta do_beta(): $f = CODE(0x7c6ee8), $_ = beta do_beta(): $f = CODE(0x7c6ee8), $_ = beta do_beta(): $f = CODE(0x7c6ee8), $_ = beta [download] As you can see, the coderef changes after alpha has been encountered. (upd: as it seems, BrowserUk posted essentially the same idea while I wrote up my reply...)	[reply] [d/l] [select]
A reply falls below the community's threshold of quality. You may see it by logging in.
Re: A way to avoid repeated conditional loops by Marshall (Canon) on Aug 24, 2011 at 13:45 UTC
The below is one way: Keep things simple. Output of below is: term 'alpha' detected on input line 4 The code below is very efficient and runs fast. `#!usr/bin/perl -w use strict; while (<DATA>) { if(/\balpha\b/) { print "term 'alpha' detected on input line $.\n"; next; } .. do something here... .. sometimes better than elsif() } __DATA__ beta beta beta alpha # this line appears only once in the whole file beta beta beta` [download] if you have a lot of conditions, then perhaps something like this is appropriate: `while ( <FILE> ) { if ( /alpha/ ) { .... next; } if (/beta/) { .... next; } #here is the default code... }` [download] One of the reasons to do it that way is so that all of the conditions can be interchanged with one another. If you have an if(..)elsif(..)else(..) conditional, this hinders the ability to change the order of the conditions. At the end of the day, that approach will result in more complex code and limit the ability to re-order the code's priority of matching - don't do it. At least for number of "if" statements of about 5-6. Update: Well this is one of those situations where there just isn't a "one size fits all" solution. I tend to prefer coding with several sequential if{..} statements so that all of them are "equal". This whole coding style discussion can set off something that will last until a gallon of gas falls back to $1.60.	[reply] [d/l] [select]
Re: A way to avoid repeated conditional loops by BrowserUk (Patriarch) on Aug 24, 2011 at 19:26 UTC
You can always disable the warning if it bothers you. Of course, its only a significant gain if the alpha appears early in the file. And, the sub overhead may kill the gain, if the processing involved isn't very significant: `#! perl -slw use strict; sub doBeta(_) { print 'Beta'; } sub doit(_) { if( /alpha/ ) { print 'Alpha'; *doit = \&doBeta; return; } doBeta( $_ ); } doit( $_ ) while <DATA>; __DATA__ beta beta beta beta beta beta beta beta beta alpha beta beta beta beta beta beta beta beta beta beta beta beta` [download] Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. "Science is about questioning the status quo. Questioning authority". In the absence of evidence, opinion is indistinguishable from prejudice.	[reply] [d/l]
Re: A way to avoid repeated conditional loops by Not_a_Number (Prior) on Aug 24, 2011 at 19:08 UTC
Hi. Frankly, if your dataset is as described in your OP, of all the solutions posted above, I prefer your original one! It has two major advantages over many of the other solutions provided: 1. It works! 2 . It is extremely comprehensible.	[reply]
Re: A way to avoid repeated conditional loops by TomDLux (Vicar) on Aug 24, 2011 at 19:25 UTC
`use autodie; open my $file, '<', $filepath; process_file_header( $file ); proress_file_remainder( $file );` [download] As Occam said: Entia non sunt multiplicanda praeter necessitatem.	[reply] [d/l]
Re: A way to avoid repeated conditional loops by locked_user sundialsvc4 (Abbot) on Aug 24, 2011 at 12:22 UTC
Call it “a flag variable” if you want to, but this is a classic place for a finite-state machine (FSM) algorithm. I presume that you are familiar with the term. Edit: One subtle distinction that sometimes comes in handy with FSM algorithms is to put “read the next line” as a subroutine-call that occurs within each state-handler, instead of structuring the algorithm within a while-loop as shown above. Sometimes you might want to loop through more than one consecutive state while processing the same line.
Re: A way to avoid repeated conditional loops by Anonymous Monk on Aug 25, 2011 at 10:20 UTC
Use m?alpha? instead of m/alpha/ to match only once: `#!/usr/bin/perl use warnings; use strict; while (<DATA>) { if (m?alpha?) { print "+ $_"; } else { print "- $_"; } } __DATA__ beta beta alpha beta beta alpha alpha beta` [download] Results in: `- beta - beta + alpha - beta - beta - alpha - alpha - beta` [download]	[reply] [d/l] [select]
Re^2: A way to avoid repeated conditional loops by Deus Ex (Scribe) on Aug 25, 2011 at 13:13 UTC
It would be a nice trick, but the match is just ignored after the first time, so the conditions keeps getting evaluated. Thanks though	[reply]
Re^3: A way to avoid repeated conditional loops by pvaldes (Chaplain) on Aug 26, 2011 at 00:45 UTC
`=~ s/^alpha/^#alpha/` ... and now do what you want with the whole file	[reply] [d/l]


Think about Loose Coupling
	PerlMonks