Shell Programming and Scripting

Blasted data inputters they always have to screw my data up....My comma delimited file with three fields ( firstname,surname and address ) has been screwed up by people entering addresses like this (putting a comma in between the house number and the street name)

142,Stonewall Avenue

We explicitly instructed them not to do this and yet you always get one dont you

So my three field comma delimited file is now a four field one on some lines eg

Gary,Jones,12 Example Street
Peter,Wilson,15,Demo Street

Im not sure theres any way to rectify this ? , maybe theres something you all could suggest.....any help would be greatly appreciated

if this is the only problem, i.e. there is only one extra comma in some records, a simple awk solution

but I am sure there could be some other complexities involved as well, if that is the case specify them,
Gaurav

yes you are right, it is more complex as but there are fields after the address field (in fact another 10 or so) . I assume your code has concatenated $3 and $4 and I agree that if they were the last fields in the record then this wouldnt be a problem, but they arent, so concatenating them would put all the subsequent fields out of line...apologies..I should have given you the full picture from the offset

Im thinking that I may have to change my delimiting character

can you give the full picture now it might help. Exactly how many fields are there and give some sample data if possible. And is it possible that no. of fields in two records may differ.
Maybe you wont have to change the delimiting character.

Try this.

Code:

#! /usr/local/bin/perl -w

open (TXT,"<input.txt");

while ($line=<TXT>) {
        if ( $line =~ m/([^,]*,[^,]*,)(.*)/ ) {
                $L = $1;
                $R = $2;
                $R =~ s/,/ /g;
                print "$L$R\n";
        }
}

Would give output as

Gary,Jones,12 Example Street
Peter,Wilson,15 Demo Street

ok, one of the deveolpers of the system that outputs the source file has just popped over to my desk saying theyre gonna change the delimiter to a * so all my problems have now been solved

thanks all for your input

Until someone uses a * somewhere in the data.

| <- pipe character -- a better user created data delimiter because people don't know/seldom use it.

It's sort of a defacto unix delimiter for this kind of thing. More of a workaround to bypass the standard idiot user, you might say.