UNIX for Advanced & Expert Users

I am trying to rename multiple files from lower case to upper case, but then I need to change the word SHIP to Ship and CSV to csv. This is a sample of the input file name input_file_ship_1234_20110106.csv I need this to look like this INPUT_FILE_Ship_1234_20110106.csv this is the result when I run the following code INPUT_FILE_Ship*Ship.csv.

# rename lower case name to UPPER CASE NAME
cat $SRC_FILE | tr '[:lower:]' '[:upper:]' > $UC_FILEA

# rename SHIP to Ship and CSV to csv
for i in `ls INPUT_FILE_SHIP*`
do
xx=`echo $i | sed -e 's/SHIP/Ship/g' | sed -e 's/CSV/csv/g'`
done

Thank you

Use code tags or get dunned by a moderator!

Really, it is better form to step up to writing a ksh script in sed from an ls. You can inspect the script before running it.

Code:

 
ls ... | sed '
  s/.*/mv & &/
  s/\(.*\)SHIP/\1Ship/
  s/\(.*\)CSV/\1csv/
  /mv \(.*\) \1$/d
 ' >/tmp/new.sh

I made sure there was a real rename possible. This pragma is very robust for large scale. You can do a whole tree with 'find ...' for 'ls ...'. You can swap '|ksh' for '>/tmp/new,sh' when you are sure it is stable.

Thank you. Actually it's my tr that is causing me issues. After adding you code and retesting getting the exact same results I removerd the sed command and just ran with the tr, I see now that the tr is causing my file to have an "*" in the name.

The sed y command can do tr sorts of things. However, case is a lower-functionality area in sed and regex, IMHO. Being a character transformer, tr is very robust except for null, since the transform arguments are null terminated strings.