UNIX for Dummies Questions & Answers

Hi.

This is my first post,

I am trying to change certain sections of a file from UPPERCASE to lowercase. For this I am using the tr command. The thing is it is changing all of the text in the file to lowercase. My question is "can you use pattern matching or something in conjunction with tr to selectively change text?"

Eg: my code is

line_x (P) { ....
line_y (Q) { ...........
line_z(P Q) {..............

I only want to make whatever is in line_y to lowercase.
(there are multiple instances of line y)

Currently I am doing something like this:

tr "*(line_y ([A-Z]) {)" "*(line_y ([a-z]) {)" < file.orig > file.mod

but this is still just making all text lower case.

Any help or input would be greatly appreciated.

Kind Regards,

Colum.

No. tr is NOT the right tool for this. You can do this sort of thing with sed (with some difficulty), awk or perl though. If you don't need this to be fast, you could also read each line, slice it up and use tr on the part that needs to be converted and reassemble the pieces.

Code:

sed "/_y/y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/" file

Code:

perl -ne ' { if( /_y/ ) { $_ =~ tr/[A-Z]/[a-z]/ } print } ' file

Code:

>echo line_y ABC DEF GHI | gawk '{print $1,tolower($2),tolower($3),tolower($4)}'

would convert the 2nd, 3rd, 4th fields on a line to lowercase

sed -e '/line_y/s/[A-Z]/\L&/g' file

It is not entirely clear what you mean. This will turn only the part in parentheses after line_y into lowercase if you are using GNU sed:

Code:

sed 's/\(line_y *(\)\([^)]*\)/\1\L\2/g' file

Thanks lads. Will play around with these solutions.