regex escape the comma

casperdaghost · 08-12-2010, 09:43 PM

sometimes there are one, sometime there are two exchanges in this log file. the 100= is the stock exchange- if there are two, they are seperated by a comma.
i understand how to \ escape a comma in a regex, but I am having trouble with combining it.

35=8 39=1 38=1000 32=13 14=754 31=1.11 44=1.1 100=AMEX,ISE
35=8 39=1 38=1000 32=151 14=205 31=1.1 44=1.1 100=AMEX
35=U 39=2 38=1000 32=45 14=184 31=1.1 44=1.1 100=AMEX,ISE
35=U 39=2 38=1000 32=45 14=574 31=1.1 44=1.1 100=AMEX
35=U 39=2 38=1000 32=45 14=734 31=1.1 44=1.1 100=AMEX,ISE
35=U 39=2 38=1000 32=45 14=704 31=1.1 44=1.1 100=AMEX,ISE
35=U 39=2 38=1000 32=45 14=794 31=1.1 44=1.1 100=AMEX
35=U 39=2 38=1000 32=45 14=484 31=1.1 44=1.1 100=AMEX
35=8 39=1 38=1000 32=0 14=905 31=0.00 44=1.1 100=AMEX,ISE
35=8 39=1 38=1000 32=44 14=949 31=1.11 44=1.1
35=U
35=8 39=4 38=1000 32=0 14=949 31=0.00 44=1.1 100=ISE

zgrep 00072224.0000f5d1.4c60d592 fixaudit.eprop323.20100810.002815.gz | perl -ne 'print /(100=[\w]*\,[\w]* )/ | sort
100=AMEX,ISE
100=AMEX,ISE
100=AMEX,ISE
100=AMEX,ISE
100=AMEX,ISE
zgrep 00072224.0000f5d1.4c60d592 fixaudit.eprop323.20100810.002815.gz | perl -ne 'print /(100=[\w]*)/ | sort
100=AMEX
100=AMEX
100=AMEX
100=AMEX
100=AMEX
100=AMEX
100=AMEX
100=AMEX
100=AMEX
100=ISE

What I want is

100=AMEX,ISE
100=AMEX
100=AMEX.ISE
100=AMEX
100=AMEX,ISE
100=AMEX,ISE
100=AMEX
100=AMEX
100=AMEX,ISE
100=ISE

Tinkster · 08-12-2010, 09:48 PM

What makes you think a comma needs escaping? It has no special meaning
in any regex flavour I'm aware of.

Anyway ... I think what you want is:

Code:

zgrep 00072224.0000f5d1.4c60d592 fixaudit.eprop323.20100810.002815.gz | perl -ne 'print /(100=[\w]+,*[\w]*)/ | sort

GrapefruiTgirl · 08-12-2010, 09:52 PM

Does it have to use perl?

Code:

awk '/100=/{print $NF}'

And why were you running it through `sort`? the result you say you want, is not sorted.

GrapefruiTgirl · 08-12-2010, 09:55 PM

And - as I'm not a perlite, I'm curious: are you two both missing a single quote, or should that work for me with only one single quote (because it doesn't)?

chrism01 · 08-12-2010, 10:03 PM

Code:

@arr=split(/\s+/, $rec);
if( scalar(@arr) == 8 )
{  
    print "$arr[7]\n";
}

Tinkster · 08-12-2010, 11:56 PM

Quote:

Originally Posted by GrapefruiTgirl

And - as I'm not a perlite, I'm curious: are you two both missing a single quote, or should that work for me with only one single quote (because it doesn't)?

You're quite right :}

I just copied & pasted and changed the regex

Tinkster · 08-12-2010, 11:57 PM

Quote:

Originally Posted by GrapefruiTgirl

Does it have to use perl?

Code:

awk '/100=/{print $NF}'

And why were you running it through `sort`? the result you say you want, is not sorted.

But that would (as 100 is 1000) also print

Code:

35=8 39=1 38=1000 32=44 14=949 31=1.11 44=1.1

Cheers,
Tink

grail · 08-13-2010, 02:01 AM

Quote:

Originally Posted by Tinkster

But that would (as 100 is 1000) also print

Maybe if the 1000 had an equals sign (=) inserted as the second last character.

Tinkster · 08-13-2010, 11:49 PM

Right you are, and I am blind.

Apologies, Sasha!

GrapefruiTgirl · 08-14-2010, 06:02 AM

Ehh, no worries Tink :-) there was a printf statement I had been using way back sometime in some code or other (when I didn't really know what printf even was) and it actually made use of a single single-quote, deliberately.. So, there are cases.

Anyhow, I'm still curious about the `sort` command the OP uses here, since his desired output is in the same order as they were in the input; maybe if the OP could comment on what everyone's posted here? Does any of this work as desired?

casperdaghost · 09-12-2010, 01:08 PM

I use the sort because the log file is 10,000 lines and it prints out the answers line by line - the sort makes sure that they are at the end of the command prompt - not spaced out over a few screens.

casperdaghost · 09-12-2010, 01:09 PM

Quote:

Originally Posted by Tinkster

Right you are, and I am blind.

Apologies, Sasha!

THanks tink