perl: Split a string?

[Evan]

Im stuck trying to split this string

It must split all of my examples. I want to seperate them by the character by these three characters

?:;


1?Not Insured:Insured;
1?No:Yes;
1?option1:option2;

[CornedBee]

Use a regular expression.

$input =~ /^(.+)\?(.+):(.+);$/
// $1, $2 and $3 hold the parts

I hope I got string begin and end right (^ and $ I think, but I'm not sure).

[Evan]

You have been some help, but that code doesnt do what I need. So let me try again.

Say I have a html page in a scalar. $html
in this page could be the string 1?Not Insured:Insured;

So I want to go threw the HTML and loop threw and put all of these expressions into 3 strings

$inp1 = 1;
$inp2 = "Not Insured";
$inp3 = "Insured";


Can you help me? Thanks so much!

[CornedBee]

What does $html contain? Some random HTML code, with this stuff embedded? Or only such triples? What more do we know about these triples except that they are separated by ? and : and finished by ;? Are they all on their own line? Is the first part always a number?

[Evan]

Its an entire html page, and these are embedded in this page.

They ALWAYS are in this format


number?option1:option2;


number is going to be which option will show up, and im gonna erase the rest of the string in the html. Get it?

So if its 1?No:Yes; it will say Yes
and 0?No:Yes; will say No

[CornedBee]

Hmmm....

There's something with execution in the second part of search-replace regular expressions that should do this nearly by itself. But I'm no expert in perl.

It's something like
$html =~ s/(\d)?(.+?):(.+?);/$1 ? $2 : $3/egs;

I believe this should work.

[Evan]

Wow!

It ate my html up like roast beef, and them spit out stuff all over the place. Funny but kinda sad!

Heh.. Any idea why it would do that?

[Evan]

$html =~ s/(\d)\?(.+?):(.+?);/$1 ? $2 : $3/egs;


You forgot to escape the ?

Thanks alot!!!

[Evan]

I made a note to my self of how you managed to do that with that regex.

Is my understanding correct?


#NOTE: This line swaps out expressions in the html that
#look like this
#num?option1?option2
#if num = 1 then uses option2, else option1

#switch / var1(digit) escape? var2(all) : var3(all) ; / if 1 then 3, else 2 / right=expression, get all, single line input
$html =~ s/(\d)\?(.+?):(.+?);/$1 ? $2 : $3/egs;

[CornedBee]

Exactly.

Except that the "all" is "one or more of any character, use the least possible"

[Evan]

So . is all char
+ is one or more, and ? is what

[CornedBee]

. means any char, + means one or more, the ? makes + non-greedy.

Consider this string:
blabla&foobar&useless
and this regular expression
(.+)&(.+)
Any character, 1 or more times, followed by &, followed by any character 1 or more times.
Now, there are two ways to match this. Way #1:
$1 = blabla&foobar
$2 = useless
Way #2:
$1 = blabla
$2 = foobar&useless

Which should apply? The standard behavior is to be greedy, that is, to match as much as possible while still matching the expression as a whole. This means that the first .+ matches everything up to the second &, which is more than matching up to the first &, so Way #1 will be chosen by default,
The ? after the + makes it non-greedy, so that it matches as little as possible. Way #2 will be chosen for the regular expression
(.+?)&(.+)

The reason why I have these in your case is simple. If your $html looks like

Some characters1?one:other;some more characters0?other:one;and some final characters.

Then a greedy match would grab
$1 = 1
$2 = one:other;some more characters0?other
$3 = one

Which is not what you want.

[Evan]

oh, that makes so much more sense now.
I wondered what 'greedy' was.

thanks alot for the help. im sure I will be asking
you lots more later =)

[Evan]

Sometimes there is no number, and it doesnt get the string.

So is it possible if there is no number to act like its 0 ???

how would you code that?

i tried this:
$html =~ s/(\d|)\?(.+?):(.+?);/$1 ? $2 : $3/egs;

the pipe didnt help.

Thanks in advance!

[CornedBee]

$html =~ s/(\d?)\?(.+?):(.+?);/$1 ? $2 : $3/egs;

The ? here means "0 or 1 times".

[Evan]

so where did you learn the regex so well, I do for work
everyday ,so I want to learn it well.


Whats the best way?

[CornedBee]

No idea actually. I read through a few tutorials and simply used them.

I found the JavaScript Regex reference by Netscape to be quite good, or the perl reference (manpages).

If you've got some *nix system, just type man perl and check which chapter you need.

On Windows, there might have come a reference with your Perl installation.