topic Re: scripting question in Operating System - HP-UX

scripting question

John McDen — Mon, 10 Jun 2002 12:44:51 GMT

How do I extract a particular pattern followed by few charaters. I have have a very large data file. There are few line containg a typical pattern I just want to extract it.

The pattern is randomly placed in the file,but has a common 1st two chars. followed by 8 diff. chars.

eg.
INPUT DATA
This is a test data XYxxxxxxxx
Data XYmmmmm_mm could be any where in the file
Need to only extract XYaaaaaaaa

OUTPUT

XYxxxxxxxx
XYmmmmm_mm
XYaaaaaaaa
....

Is it possible to do this ?? Please help I am brain dead ... can't think of anything..really.

Re: scripting question

Rodney Hills — Mon, 10 Jun 2002 12:54:12 GMT

A simple perl program can do it

open(INP,"while() {
chomp;
while (/(XY[^\s]*)/g) {
print $1,"\n";
}
}

-- Rod Hills

Re: scripting question

John McDen — Mon, 10 Jun 2002 13:20:24 GMT

Rodney thanks for replying so fast it works but I just need the XY and next 8 chars I wish I knew perl so that I could fix it on my own.

Re: scripting question

Rodney Hills — Mon, 10 Jun 2002 13:23:30 GMT

Just change

while (/(XY[^\s]*)/g) {

to

while (/(XY........)/g) {

In your sample, you had a value with _mm appended which made it more than 8 characters so I wasn't sure if you needed the characters up to the next blank character or end of line.

Good Luck

-- Rod Hills

Re: scripting question

Rodney Hills — Mon, 10 Jun 2002 13:27:53 GMT

Oops, recounting I see it was 8 characters. These variable width fonts can be misleading at times.

-- Rod Hills

Re: scripting question

John McDen — Mon, 10 Jun 2002 13:28:30 GMT

Great .. it worked... Thanks

Re: scripting question

harry d brown jr — Mon, 10 Jun 2002 13:52:36 GMT

Or in a shell:

cat DATA | tr -cs "[:alnum:][:punct:]" "[\012*]" | grep XY

live free or die
harry

Re: scripting question

Ralph Grothe — Mon, 10 Jun 2002 13:52:38 GMT

perl -ne 'print "$1\n" if /\b(XY\w{8})/' /path/to/your/file

Re: scripting question

Rodney Hills — Mon, 10 Jun 2002 13:57:05 GMT

Ralph, I believe your version assumes only one per line. The version I supplied allows for multiple entries to exist on a line.

See documentation on m//g. "g" allows to search multiple times through the same line.

-- Rod Hills

Re: scripting question

harry d brown jr — Mon, 10 Jun 2002 14:03:05 GMT

John,

This will hit those "words" starting with XY with only a length of eight:

cat DATA | tr -cs "[:alnum:][:punct:]" "[\012*]" | grep "^XY$.$\{8\}$"

ignoring anything like XYabcdefghijk

live free or die
harry

Re: scripting question

John McDen — Mon, 10 Jun 2002 14:11:52 GMT

Thanks guys Rodney's solution was perfect...I really appreciate everybody help ..

Re: scripting question

H.Merijn Brand (procura — Mon, 10 Jun 2002 14:24:34 GMT

Rodney is going to steal my perl answers? :]}

N/A

Personally I thing Ralph's ansewr is better