Awk - substitution with parts of the matched regular expression used.

c_moriarty · 04-30-2011, 08:25 AM

I remember reading that using sed, you can do this with parentheses:
s/abc(something)def/(something)else/g

I can't find an explanation of how to do something like this with Awk.
Say you have this in an HTML file, where (number) stands for a one or two-digit number:
(number)

And you want to change that to this:
<a name="(number)t" href="#(number)b">(number)</a>

How would this be possible? Would you have to use sed?

c_moriarty · 04-30-2011, 08:29 AM

This is what I was going to do, exactly, with the pages of "Practical Common Lisp":

Original:
(number)
Modified:
<a name="(number)t" href="#(number)b">[(number)]</a>

The brackets [] I put there to change the 'd number to the 'd number in brackets. From (number) to [(number)].

colucix · 04-30-2011, 08:32 AM

GNU awk provides the gensub function. Example:

Code:

awk '{print gensub(/<sup>([0-9]+)<\/sup>/,"<a name=\"\\1t\" href=\"#\\1b\"><sup>\\1</sup></a>","g",$0)}' file

where the reference to the parenthesized field in the regexp is \\1 in the replacement string. See http://www.gnu.org/software/gawk/man...007d_0029-1200 for details.

c_moriarty · 04-30-2011, 08:36 AM

Most specifically, what I was going to do is this:

There are two occurences of (number)

One earlier in the file, one later in the file.
I was going to make these changes to the first one:
From (number) to <a name="(number)t" href="#(number)b">[(number)]</a>
and these changes to the second one:
From (number) to <a name="(number)b" href="#(number)t">[(number)]</a>

to create a clickable link between the two, so you don't have to scroll to the bottom to see the notes and back to where you were again to continue reading where you left off.
But this is so generally useful that I would really like to know how to do this.

c_moriarty · 04-30-2011, 08:38 AM

Thank you.