[ILUG] sed advice required...

[Kevin O'Riordan]

> i.e. text_I_want_to_remove=1234=part of string I want
> and text_I_want_to_remove=12345=part of string I want
> 
> ... what I have at the moment is
> 
> sed -e 's/text_I_want_to_remove=[0-9][0-9][0-9][0-9]=//' file ...


The sed regex:

    \([0-9]\+\)

would match and capture a string of length 1 or more consisting of chars 0-9
or

    \([0-9]*\)

would match and capture a string of length 0 or more consisting of chars 0-9.
or

    \([0-9]\{4,6\}\)

would match and capture a string of length 4 to 6 consisting of chars 0-9.



I'm not sure from your example text, but you might be able to match
portions based on a delimiter, such as '='.  If your lines were to be
of the form:

    someJunkText=12345=someImportantText

then you could match strings of chars that don't include the delimiter,
separated by the delimiter, like so:

    sed -e 's/^\([^=]\+\)=\([^=]\+\)=\(.*\)$/\3/'

The carets in brackets here negate the range of chars in the brackets
...  the regex(7) manpage will explain this a helluva lot better than
I will.  The only issue then is where to use backslashes to escape
particular special chars such as parentheses, addition-signs, etc.
Every regex-using program - such as perl, sed, vi, or grep - differs
here.


hth,
-kev


-- 
You can watch for your administrator to install the latest kernel with
	watch uname -r

		-from the manpage for watch(1).