Perl - need regex help

[iDxMan]

I've suddenly gone brain dead. -> Please help! :D

Here's what I need out of this string.

sample data:

Identifier: Some extra crap in the line: More crap: STANZA1=WORD1 WORD2 WORD3 STANZA2=STUFF123 STANZA3=MORE STUFF STANZA4=4327780917801 STANZA5=03/01/2001


I need words1-3 in stanza1 -- although there might be 1 or 2 at times.

The stanza's are the same, but they aren't always there so I can't depend on #2 being there in order to grab everything in between.

There are only 5 possibilities for the stanza's so I could include them all, but who wants that?

So far this works *IF* #2 is there:


($var) = /.+STANZA1=(.+)STANZA2=/;


so $var is "WORD1 WORD2 WORD3"

If #2 isn't there I don't get the data. In various other tests I get the data, but the next stanza name is in there too.

Any ideas?

Thanks...

[Mikey123]

how about:

(@match)= split(/STANZA\d+=/);


it leaves you a null first element but splits it up correctly after that i believe.

[ 01 May 2001: Message edited by: Mikey123 ]

[iDxMan]

That would work if the stanza's were actually named accordingly. ack! Guess I should have mentioned that..

They are something like:

NAME=
DOB=
SSN=
PROVIDER ID=
MEMBER ID=

The NAME= always seems to be first and that's the one I need. So far, depending on the situation, either MEMBER ID or SSN is next..

-r

[ 02 May 2001: Message edited by: iDxMan ]

[iDxMan]

This seems to work, but I would rather have it look for the next stanza after NAME -- not hard code ssn and member id in.. Who knows if the return data will change..


($var) = /.+NAME=(.+)(SSN=|MEMBER ID=)/;

[YaRness]

Originally posted by iDxMan:
<STRONG>
Identifier: Some extra crap in the line: More crap: STANZA1=WORD1 WORD2 WORD3 STANZA2=STUFF123 STANZA3=MORE STUFF STANZA4=4327780917801 STANZA5=03/01/2001


I need words1-3 in stanza1</STRONG>

did this with perl:

$_ = 'Identifier: Some extra crap in the line: More crap: STANZA1=WORD1 WORD2 WORD3 STANZA2=STUFF123...';
/STANZA1=((?:[^\s]+\s+){1,3})/ and $var = $1;

print $var;

that will match if there is at least one word (a word in this case being any non-whitespace characters followed by one or more whitespaces) in STANZA1, and will pull out up to three words.

[iDxMan]

Thanks Yarness... &lt;&lt; But its a bit too greedy.


$_ = 'Identifier: Some extra crap in the line: More crap: STANZA1=WORD1 WORD2 STANZA2=STUFF123...';
/STANZA1=((?:[^\s]+\s+){1,3})/ and $var = $1;


Now $var = "WORD1 WORD2 STANZA2"

or if stanza1 only has 1 value then $var is "WORD1 STANZA2=STUFF123"

One other problem is that stanza2 can be in the form of:
STANZA2=DATA
or
STANZA2 ID=DATA

I'll play around with that for a bit.. Its a start. :D

-r

[ 02 May 2001: Message edited by: iDxMan ]

[Mikey123]

(@match)= split(/\w+=/);
shift @match;

split on a series of word characters followed by an equals sign then shift off the first empty element. If there is a space in your identifier you will have to modify the above regex a little. It might be easier to ensure you don't use spaces ie Member_ID= instead of Member ID

Thinking a little more about the "MEMBER ID" thing I think you will either have to ensure no spaces are used for identifiers or seperate each block with a comma or something. Otherwise there is no way of knowing if MEMBER is part of the previous stanza or part of the upcoming identifier.

[ 02 May 2001: Message edited by: Mikey123 ]

[YaRness]

yup, sure did goof that one up. replace the regex with

/STANZA1=((?:[^\s]+\s+)+)STANZA2=/

it will get all the words, or none if there are none (tested it too!). you can substitute the bolded part here, ((?:[^\s]+\s+)+), for something like {1,3} if you want between 1 and 3 words, or {2,} i think if you want 2 or more words. see the perlre man page for more info ('perldoc perlre' or 'man perlre', or browse to it through your activestate documentation... whichever works for ya).

[iDxMan]

Originally posted by Mikey123:
<STRONG>
(@match)= split(/\w+=/);
shift @match;

split on a series of word characters followed by an equals sign then shift off the first empty element. If there is a space in your identifier you will have to modify the above regex a little. It might be easier to ensure you don't use spaces ie Member_ID= instead of Member ID
</STRONG>

Thanks for the ideas.. I'll poke around..

For other parts of the data I have been converting space to an underscore for ease of use.. But this part is a bit different.

data:
Date of Birth: 01/01/1970

storage:
$key{DATE_OF_BIRTH} = "01/01/1970";


-r

[Mikey123]

I edited my previous post to say this but if you do not either delimit the blocks or use _ for items like MEMBER_ID it will be impossible to know if "previousdata Stanza2" is an identifier or if "previousdata" belongs to the previous group.

[iDxMan]

Originally posted by Mikey123:
<STRONG>Thinking a little more about the "MEMBER ID" thing I think you will either have to ensure no spaces are used for identifiers or seperate each block with a comma or something. Otherwise there is no way of knowing if MEMBER is part of the previous stanza or part of the upcoming identifier.
</STRONG>

Exactly.. Now the real fun begins when they start to change the format of the return data. :)

(`they` being the company who returns result data for us to update our system with.)

Boy I love last minute hacks due to new software that doesn't have the same features as the old version. lol

-r

[iDxMan]

Old post, but I finally found time to look at it again, so here's an update.. This seems to work best so far..

I kept running into the problem of matching one but not the other.

ie:

NAME=FIRST MID LAST STANZA2=THIS
but not
NAME=FIRST MID LAST STANZA1 IDNUM=THIS

or it being too greedy..

ie: NAME=FIRST LAST STANZA1 IDNUM=THIS
would grab "FIRST LAST STANZA1" as the value..

anyways:

Probably could combine this into 1 regex, but this will have to do..

-r


if(/.+NAME=((\w+\s){1,3})(\w+\s\w+\=)/)
{
($name) = $1;
} else {
($name) = /.+NAME=((\w+\s){1,3})(\w+\=)/;
}