AWK - Re: confused with regular expression

Author

Re: confused with regular expression

Ed Morton

2004-11-16, 6:50 pm

moggces wrote:
> Sorry
> I am not sure what was wrong.
>
> I have a file which has two pattern
>
> 5673058-
> AE-848/14271200-
>
> in command line I typed:
> awk '/^[0-9]/ { print }' file
> , tried to get "5673058-" but it printed nothing
> so I tried to type
> awk --re-interval '/[0-9]{7}/ { print}' file
> strangely, it didn't print "5673058-" but instead it printed "AE-848/14271200-"
>
> my gawk version is 3.1.4

Mine is 3.1.3 and it works as expected:

$ printf "5673058-\nAE-848/14271200-\n" | awk '/^[0-9]/ { print }'
5673058-

$ printf "5673058-\nAE-848/14271200-\n" | awk --re-interval '/[0-9]{7}/
{ print }'
5673058-
AE-848/14271200-

$ awk --version
GNU Awk 3.1.3

So, at least you know it isn't your RE(s). If grep or sed don't find
them, it's your file that's got control chars or something, otherwise
it's your gawk version.

Regards,

Ed.

moggces

2004-11-16, 6:50 pm

arnold@skeeve.com (Aharon Robbins) wrote in message news:<4188a048@news.012.net.il>...
> In article <BJ6dnfoEnfzB7xXcRVn-gw@comcast.com>,
> Ed Morton <morton@lsupcaemnt.com> wrote:
> "AE-848/14271200-"
>
> Actually, I'm 99% certain it's not the gawk version, but the locale; the
> original poster is in China. It's probably better to use
>
> /^[[:digit:]]/
>
> and
>
> /^[[:digit:]]{7}/
>
> instead of [0-9]. I suspect that if the original poster uses
>
> export LC_ALL=C
>
> before running gawk, things will work as expected.
>
> Arnold

It didn't work when I use /^[[:digit:]]/ without "export LC_ALL=C"

and both /^[0-9]/ and /^[[:digit:]]/ could work after "export LC_ALL=C"

Thanks very much for all responses

Jui-Hua
from Taiwan

Sponsored Links