Re: [Sheflug] Replacing a block of text in a file with text fromanother file

[Jens Thomas <j.m.h.thomas [at] xxxxxxxx>]

Thanks for the tips David and Chris - I really appreciate you taking the 
time to give me your suggestions.

I fiddled around a bit more and have managed to come up with a 
work-around, although it looks pretty horrible, and I'm sure there's a 
neater way to do it .

#!/bin/bash
sed -e '/zmatrix angstrom/,/end/cREPLACEME' <$1 >$1.tmp
sed '/REPLACEME/ {
r basis.txt
d
}' <$1.tmp >$1._in
rm $1.tmp

The link to the sed faq that David fowarded was really useful and I also 
found the following pretty good too, in case anyone else wants to delve 
into sed (something I wouldn't recommend unless you are pretty 
bloody-minded or seriously averse to Perl or Python).

http://www.grymoire.com/Unix/Sed.html
http://www-h.eng.cam.ac.uk/help/tpl/unix/sed.html

Thanks again,

Jens

Chris Johnson wrote:

> > I'm trying to write a script that will go through a bunch of
> > input files
> > I have and replace all the text between the lines 'zmatrix
> > angstrom' and
> > 'end' with text the from a file (the input file is test.in
> > and the text
> > to be inserted from basis.txt in  the example below).
> >    
>
> I'd use (and have used) perl to do this.  I had a template html file that
> needed stats inserting at a certain point.
>
> Not sure how good your perl is (hopefully better than mine) but the
> structure should look a bit like this.
>
> Read inputfile line
> if inputfile line = starttag then {
> 	open insertfile
> 	read insertfile line by line
> 	write insert file lines to outputfile
>
> 	read input file until inputfile line = endtag #(this will skip to the end
> of the bit to miss out)
>      }
> if inputfile line <> start tag and inputfile line <> end tag write inputfile
> line to outputfile.
>
> Hope this isn't too crypic.  I don't have time at the mo to put it in real
> code and without testing I'd probably mix my php and perl together and leave
> you with a load of useless code.
>
> It can probably be done in a lot less code.  You have to get perl/sed etc to
> match over multiple lines which I was never any good at and it took me less
> time to code it in long hand than try to be clever and do it in a one liner.
>
> HTH
>
> Chris Johnson
> (No not that one, the other one!)
> (running SUSE 9, 512MB PIII 600 (ish))
>
> ---
> Outgoing mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.734 / Virus Database: 488 - Release Date: 04/08/2004
>
>
> ___________________________________________________________________
>
> Sheffield Linux User's Group -
> http://www.sheflug.co.uk/mailfaq.html
>
>  GNU the choice of a complete generation.
>
>  
___________________________________________________________________

Sheffield Linux User's Group -
http://www.sheflug.co.uk/mailfaq.html

 GNU the choice of a complete generation.