I want to use regexp to match a string, say "ABC", that is either
preceeded by "X" or is at the beginning of the text string.  How can I
create a regexp that conditionally checks for X or ^ (line anchor)?

[ Post a follow-up to this message ]

nonsequitur wrote:
> I want to use regexp to match a string, say "ABC", that is either
> preceeded by "X" or is at the beginning of the text string.  How can
I
> create a regexp that conditionally checks for X or ^ (line anchor)?

How about '(^|.*X)ABC' ?

jc@soyuz:/tmp$ cat test.input
ABC (should match)
zzz ABC (should not match)
yyy XABC (should match)
XABC (should match)
WABC (should not match)

jc@soyuz:/tmp$ egrep '(^|.*X)ABC' test.input
ABC (should match)
yyy XABC (should match)
XABC (should match)

[ Post a follow-up to this message ]

"shakahshakah@gmail.com" <shakahshakah@gmail.com> writes:

> nonsequitur wrote: 
> I 
>
> How about '(^|.*X)ABC' ?

That will work with some regexp implementations, but not all.  In
some variants, ^ and $ lose their special meanings inside parens.
Also, don't forget that some variants (e.g. emacs) use \(\) as
subpatter delimiters.

--
Måns Rullgård
mru@inprovide.com

[ Post a follow-up to this message ]

M=E5ns Rullg=E5rd wrote:
> "shakahshakah@gmail.com" <shakahshakah@gmail.com> writes:
> 
can[vbcol=seagreen] 
anchor)?[vbcol=seagreen] 
>
> That will work with some regexp implementations, but not all.  In
> some variants, ^ and $ lose their special meanings inside parens.
> Also, don't forget that some variants (e.g. emacs) use \(\) as
> subpatter delimiters.
>
> --
> M=E5ns Rullg=E5rd
> mru@inprovide.com

You can always do '^(ABC|.*XABC)' then (though the subpattern
delimiters issues is still there, I guess).

[ Post a follow-up to this message ]

"shakahshakah@gmail.com" <shakahshakah@gmail.com> writes:

> Måns Rullgård wrote: 
>
> You can always do '^(ABC|.*XABC)' then (though the subpattern
> delimiters issues is still there, I guess).

^(|.*X)ABC is less repetitive.

--
Måns Rullgård
mru@inprovide.com

[ Post a follow-up to this message ]

M=E5ns Rullg=E5rd wrote:
> "shakahshakah@gmail.com" <shakahshakah@gmail.com> writes:
> 
for[vbcol=seagreen] 
>
> ^(|.*X)ABC is less repetitive.
>
> --
> M=E5ns Rullg=E5rd
> mru@inprovide.com

Interesting, I didn't know you could have an "empty" portion of a
parenthesized subexpression. You learn something new every day...

[ Post a follow-up to this message ]

nonsequitur wrote:
> I want to use regexp to match a string, say "ABC", that is either
> preceeded by "X" or is at the beginning of the text string.  How can I
> create a regexp that conditionally checks for X or ^ (line anchor)?

Try:

^X?ABC

--
Thomas M. Sommers -- tms@nj.net -- AB2SB

[ Post a follow-up to this message ]

"T.M. Sommers" <tms@nj.net> writes:

> nonsequitur wrote: 
>
> Try:
>
> ^X?ABC

Counter-example: "abcXABC".

The OP could try ^(|.*X)ABC if he's using extended regular expressions
(such as used by egrep(1)).

mkb.

[ Post a follow-up to this message ]

Matthias Buelow <mkb@incubus.de> writes:

> "T.M. Sommers" <tms@nj.net> writes:
> 
>
> Counter-example: "abcXABC".
>
> The OP could try ^(|.*X)ABC if he's using extended regular expressions
> (such as used by egrep(1)).

^(.*X)?ABC is another variation.

--
Måns Rullgård
mru@inprovide.com

[ Post a follow-up to this message ]

Matthias Buelow wrote:
> "T.M. Sommers" <tms@nj.net> writes:
> 
>
> Counter-example: "abcXABC".
>
> The OP could try ^(|.*X)ABC if he's using extended regular expressions
> (such as used by egrep(1)).

I read it to mean that the X was alone at the beginning of the
line.  Rereading indicates I was wrong in that understanding.

--
Thomas M. Sommers -- tms@nj.net -- AB2SB

[ Post a follow-up to this message ]