Re: [edlug] Regular Expression Substitution

[Martin Fraser <martin@xxx.xxx.xxx>]

The value in file 1 will more than likely not be unique. There is no 
comparison between the files. All I know so far is that the first record 
in one file should be substituted in the other.

I had thought about generating matches fro the second file, then running 
sed over the first to substitute in those matches, which would be a two 
pass approach, but I'm not sure how to get sed to do that in practice.

Thanks for the suggestion though.

mdf...

Andrew Neil Ramage wrote:
> Why not do it in two stages ?  First generate a sed file from file 2 :
>
> {
>    s/file1.value/file2.value;
>    ...
> }
>
> then run sed on file 1 using this file as the input ?
>
> Andrew
>
> Martin Fraser wrote:
> > Thanks to those that answered. I've been fighting on with this so 
> > I'll clarify what I'm up to.
> >
> > We have two files, outputted from different processes. File 2 
> > contains the result of processing File 1. Each file contains position 
> > data in the form of Eastings and Northings, i.e.  0000.00N, 0000.00E
> >
> > File 2 has slightly different (more accurate) position data, and our 
> > client wants the File1 positions replaced by the File2 positions in 
> > File 1 so they can do further processing of that data.
> >
> > The small binary headers to each line in File 1 are not a difficult 
> > problem, they are well specified and won't match any regular 
> > expression that gets the bits of the line I need, but since the line 
> > start markers are in binary, there are no carriage returns of line 
> > feeds in the file so a line by line comparison can't be done.
> >
> > File 2 contains loads of extra information but each line that I'm  
> > interesting in starts with the letter S. I'm assuming there is going 
> > to be a one to one correlation between the two files so the first 
> > match in File 1 should be substituted for the first match in File 2 etc.
> >
> > I've already written the (very complicated) regexp that matches the 
> > easting and northings in File 1 and, using sed, can substitute the 
> > values for place holders NNNNNN.00N and EEEEEEE.00E for example, so I 
> > can see in the output that I have matched the correct parts of the 
> > line. It is working perfectly, but I can't see a way to make the 
> > regexp substitutions come from a different file, or an array lookup 
> > or anything that isn't just the static replacements with back 
> > references.
> >
> > I hope this makes things clearer and some perl or regexp guru out 
> > there can help. This is my first serious use of regular expressions 
> > so I might be missing something obvious, any push in the right 
> > direction will be much appreciated.
> >
> > Cheers.
> >
> > Martin...
> >
> > -
> > ----------------------------------------------------------------------
> > You can find the EdLUG mailing list FAQ list at:
> > http://www.edlug.org.uk/list_faq.html
>
> -
> ----------------------------------------------------------------------
> You can find the EdLUG mailing list FAQ list at:
> http://www.edlug.org.uk/list_faq.html

-
----------------------------------------------------------------------
You can find the EdLUG mailing list FAQ list at:
http://www.edlug.org.uk/list_faq.html