replacing the first instance of a character in each line

[MAD_2]

I would like to know if someone can give me ideas on how to replace the first, and or second, or third, or fourth instance (whatever may be the case) of a charachter in a file by any methods, such as:

- awk & sed
- Perl
- vi substitution

Or some other way. Example:

"Doe, John"|"1234 Imagine St., Apt. 1530, Weston FL 33327"|"954-123-4567"|"954-987-6543"

I would like to separate the "Doe, John" in the first field and make it two fields by subsituting the comma with "|" to obtain results like "Doe"|"John"| (but I have about 500-1000 records to do this with so forget the manual change. Of course, each of these characters to be changed are located each in different positions and not a fixed column through the file.

Then I would have to separate full addresses into address1, address2, city, state, zip, etc. maybe using a similar method.

Thanks.

[H.Merijn Brand (procura · Accepted solution]

# perl -pe's/\s*,\s*/"|"/' file >newfile

Enjoy, have FUN! H.Merijn

[MAD_2]

Great procura... What in the code is telling it to substitute only the first instance in this case? (Is it the \s*, as in look at the beginning of the string to the first comma?) How about if I want it to replace the second, or third, or whatever instance of the character in each line?

Thanks!

[Stefan Farrelly]

cat file | sed 's/,/"|"/1' >newfile

By default sed only replaces the first occurance of the comma in each line, the 1 at the end means the same thing, only replace the first comma.

[MAD_2]

So basically I can tell it to replace whatever occurrence in each line, by replacing the 1 with a 2, 3, so on or so forth?

[Stefan Farrelly]

no, if you change the 1 in sed to 2 or 3 it will change the first 2 or 3 occurances (inclusive), not just the 2nd or 3rd. Change it to g to change all (g=global)

[MAD_2]

Ok, I am getting a little confused about your answer. If I change the 1, to let's say 2, will it only change the 2nd occurrence or will it change the 1st and 2nd? How can I make it:

1. Change the 1s, 2nd, and 3rd. Or just
2. Change only the 2nd or the 3rd

For example.

I don't want to do any global changes. Thanks!

[curt larson_1]

#!/usr/bin/ksh

file=yourfile
f="," # character to find
s="|" # character to substitute
o=1 # occurance of character to make substitution on

cat $file |
awk -v f="$f" -v s="$s" -v o="$o" '{
num=split($0,a,f);
if ( o >= num ) {print $0;}
else {
for ( i=1;iprintf("%s",a[i]);
if ( i == o ) printf("%s",s);
else printf("%s",f);
}
printf("\n");
}}'

[Michael Schulte zur Sur]

Hi,

here is an awk version:
awk '{sub(/,/,"|");print}' file

it changes only the first , in each line.

greetings,

Michael

[Ralph Haefner]

Ok, here's my goofy idea. Say you want to change the 3rd comma to a pipe. Couldn't you take Stefan's sed command and run it for the value of 3, which changes all of the first 3 commas to pipes. But then run it again with a value of 2 and reverse the pipe and comma, which changes the first 2 back?

That way you get only the 3rd one to change as you want, and you do it in 2 lines instead of needing a much longer script.

Primitive I know, but seems like it should work. I have no idea how doing that would compare to the scripts other people have suggested performance-wise, but with only 500-1000 records it probably doesn't matter.