Shell Programming and Scripting

Hello, I want to rename multiple files and catch some points about backreference within sed and regex.
Here is a part of my file list. Input:

Code:

S92A.fa
S92B.fa
... 
S96Z.fa
S921.fa
S922.fa
... 
S997.fa

Note: The file names are not necessarily continuous from A~Z or 921 ~ 997, as some of the files may be missing and there are other types of the file with different pattern in the same folder.
I want to put the prefix "CGW-" to each file name and change the extension to fasta. Output:

Code:

CGW-S92A.fa
CGW-S92B.fasta
... 
CGW-S96Z.fasta
CGW-S921.fasta
CGW-S922.fasta
...
CGW-S997.fasta

I have used two steps do the job, one at a time:

Code:

for i in S9*.fa; do mv $i $(echo $i | sed 's/S9/CGW-S9/g'); done

then

Code:

for i in CGW-S9*.fa; do mv $i $(echo $i | sed 's/fa/fasta/g'); done

It seems using the backreference will do the same job better and easier. But I could not find clues in this forum or Google. Any expertise is greatly appreciated. Thanks a lot!

You don't need sed for this:

Code:

for i in S9*.fa; do
    mv "$i" CGW-"$i"sta
done

When using sed, you should not use the substitution command's global flag, g, when your intent is to modify only a single occurrence of the pattern.

If it were at all possible that the portion of the filename matched by the wildcard, *, could contain "S9" or "fa", you'd need to be much more careful. Beyond dropping the global flag, to ensure that only the file extension suffix matches, you would need to anchor that pattern to the end of the string.

Since the file glob in the for-loop list has already constrained the matching filename to begin with S9 and end with .fa, the correct sed command doesn't have to bother being specific: sed 's/^/CGW-/; s/$/sta/'. If that file glob were not present, then the following could be used: sed '/^S9.*\.fa$/{s/^/CGW-/; s/$/sta/;}'.

Regards,
Alister

Thanks Alister!
Another thing in my mind is the back reference. How to do the job with back reference? Say, I want to change extensions ".fa", ".fx" or ".fq" to ".fasta"? Thanks again! Yifang