text search and replacement: bash scripting

jettachamp26 · 02-25-2008, 10:46 AM

That font check did the trick. I'm going to leave out the ^ in case there happens to be a font tag stuck somewhere in the middle of the file.

As for the redundant anchor tags, I might just have to deal with it and go back later and clean it up manually. It doesn't hurt my code... just doesn't need to be there/ is sloppy.

thanks again!

jettachamp26 · 02-25-2008, 11:59 AM

found another little gremlin in my code. I have a problem when I remove tags, if there is some other tag in between, its getting removed as well.

Example code:

Code:

<P STYLE="font-weight: medium"><SPAN STYLE="font-style: normal"><SPAN STYLE="font-weight: medium">After</SPAN></SPAN><I><SPAN STYLE="font-weight: medium">
The Lost Colony, </SPAN></I>

my replacement commands:

Code:

sed -i 's/<I>/<em>/g' "$f"
sed -i 's/<\/I>/<\/em>/g' "$f"
sed -i 's/<SPAN STYLE=".*">//g' "$f"
sed -i 's/<\/SPAN>//g' "$f"

**EDIT**

Its hard to see, but what I mean is in the first code set, the word After gets removed when the span tags are removed. The  tag gets removed as well.

**EDIT**

druuna · 02-25-2008, 12:38 PM

Hi,

This part:  is not unique enough. Looking at the example line I doubt if this could be done with a 'simple' reg-exp, there's not enough uniqueness to create one that isn't gready.

jettachamp26 · 02-25-2008, 12:47 PM

looks like the <P tag in the example i greedy too. Man this stinks.

jettachamp26 · 02-25-2008, 01:53 PM

figured out a solution to the problem. Both for SPAN and P to stop them from being so stinking greedy.

sed -i 's/<P[^>]*>//g' "$f"
sed -i 's/<SPAN[^>]*>//g' "$f"

by putting the [^>] in, it searches for everything but the > making it stop right at the end of its own tag, instead of later down the line.

chrism01 · 02-25-2008, 06:00 PM

Perl XML recommendations: http://www.perlfoundation.org/perl5/...ed_xml_modules
See also : http://www.perlfoundation.org/perl5/...d_cpan_modules
and the motherlode is here: search.cpan.org
For highly informed opinions, see the gurus at www.perlmonks.org