Sorting data based on criteria in Perl

[Jeff Colyer]

I have a data file that contains a set of "batches" that I need to grab out records that meet a certain search criteria. The catch is I need to tie the records back to the batch header. Here's an example of the data file.

batchname|BNO123|BDA2006-02-19|
REC|DNO11|PNO1111|UPR0999
REC|DNO11|PNO1114|UPR0999
REC|DNO11|PNO1116|UPR0999
batchname2|BNO124|BDA2006-02-19|
REC|DNO11|PNO1112|UPR0999
REC|DNO11|PNO1111|UPR0999
REC|DNO10|PNO1111|UPR0999

I need to be able to sort by both DNO and PNO as they aren't unique. A sort where DNO=11 and PNO=1111 should yield this results:
batchname|BNO123|BDA2006-02-19|
REC|DNO11|PNO1111|UPR0999
batchname2|BNO124|BDA2006-02-19|
REC|DNO11|PNO1111|UPR0999

I'll be passing the search parameters in through a shell script but I'm a perl newbie and am having difficulty trying to use hashes to parse the data. Thanks in advance!

[Muthukumar_5]

Simply as,

# grep -E '^batchname|DNO11\|PNO1111' filename

will give results.

Change DN011\|PNO1111 based on requirement.


--
Muthu

[Muthukumar_5]

With perl simply as,

perl -ne 'print if /^batchname|DNO11\|PNO1111/' filename


change DN and PNO part based on your requirement.

--
Muthu

[Jeff Colyer]

Sorry I should've been more clear. The batchname could start with anything. There is no standard naming convention. I need a way to associate the batch header with the record as the batch header could be associated with hundreds of records and there could also be 30 or more batches in one data file.

- Jeff

[Muthukumar_5]

oops. Change batchname to,

perl -ne 'print if /BNO[0-9]*|DNO11\|PNO1111/' filename

--
Muthu

[Jeff Colyer]

This actually listed every single batch header in the file and no records.

-Jeff

[Jeff Colyer]

I'm going to add an actual data file for clarity. It's from a unix system. The field delimeter is actually /375. I provided the above for simplification.

[Muthukumar_5]

I could not understand the file delimiter in,

050269ýBNO426823ýBDA2006-02-19ýBTI080303ýEFD2006-02-18ýEFT00:01ýSRB0ýBTYOýRNO14ýSNO100ý

/375? what it is?

--
Muthu

[Jeff Colyer]

It was late when I typed that. It's actually \375. I'm not sure why this was picked for a delimeter it's just in the data that is given to me. (I believe it has something to do with a vendor built driver) That's why I tried to simplify and use a | for the example. If it's easier you can do a sed and replace \375 with a | on the data I attached. Thanks again!

[Jeff Colyer]

Also, to see the \375 delimeter you must vi the file on a unix system. If you try to cat or open in notepad it garbles the delimeter.

-Jeff

[Rodney Hills · Accepted solution]

If you already have DNO and PNO in shell variables, then may this script will do what you are looking for-

$dno="DNO".$ENV{DNO}; $pno=PNO.$ENV{PNO};
while(<>) {
chomp;
@a=split("\375",$_);
if (substr($a[1],0,3) eq "BNO") {
$hdr=$_;
} else {
if ($a[1] eq $dno and $a[2] eq $pno) {
print $hdr,"\n";
print $_,"\n";
}
}
}

Then run with-
perl thisscript yourinputfile >resultfile

HTH

-- Rod Hills

[Jeff Colyer]

Thanks Rod, that worked perfectly.

[Jeff Colyer]

If you already have DNO and PNO in shell variables, then may this script will do what you are looking for-

$dno="DNO".$ENV{DNO}; $pno=PNO.$ENV{PNO};
while(<>) {
chomp;
@a=split("\375",$_);
if (substr($a[1],0,3) eq "BNO") {
$hdr=$_;
} else {
if ($a[1] eq $dno and $a[2] eq $pno) {
print $hdr,"\n";
print $_,"\n";
}
}
}

Then run with-
perl thisscript yourinputfile >resultfile

HTH

-- Rod Hills