Oh my... ok, let me give it a shot:

Code:

#!/bin/ksh
#%W% %I% %D% %T% ---- these are comments (starting with #)
#%W%G --- they don't do anything

The beginning of this line here:

Code:

//g'

Seems like something's missing there. To me it looks like the end of some global substitution using sed, something like

Code:

 sed 's/delete_this_pattern//g'

Furthermore, you don't have any input specified there... Where is the data, that these filters operate on, coming from? A file? Pipe? It's gonna want input from stdin...
if statement needs to be closed with 'fi'.

Now for the filters alone, they don't look extremely difficult to understand.

[CODE]egrep -v '(.sh:|.ksh'[/CODE] is gonna skip lines that contain '.sh:' alebo '.ksh:'
The second filter

Code:

sed 's/^.*://g'

will erase everything before the last colon, including the colon itself. 'g' at the end is not needed.The third filter:

Code:

sed 's/^M//g'

will erase all characters "^M". (these can be seen as delimiters in some windows files.)

Fourth:

Code:

grep -v '^[ s]*#'

will skip (not print for further processing) lines that start with a space or letter 's' present zero or more times and a pound sign. E.g. a line like this:

Code:

sss ss s#something

would be skipped. The line

Code:

sss ttt ff#blah

would not be skipped (there are other chars before pound sign). The line has to contain the pound sign, to be skipped. The next filter

Code:

egrep -i '(.dat|.ctl)'

will keep only lines with '.dat' or .ctl, case insensitive (e.g. '.cTl' would be matched)

This long filter:

Code:

Code:

sed -e 's/^.*=//g' -e 's/,/ /g' -e 's/;/ /g' -e 's/"//g'

has four parts: 1. it erases everything before an equal sign, including.
2. changes all commas to spaces
3. changes all semicolons to spaces
4. erases all double quotes.

This filter

Code:

tr ' ' '\n'

Will change (translate) all spaces into newlines

This one

Code:

sed 's/[    ]*//g'

will erase all spaces (there are none by this time)
This

Code:

 sed '/^$/d'

will erase all empty lines.
And finally,

Code:

sort -u | nawk -F"/" '{print $NF}'

will sort the output lines, printing only the first occurence, if there are multiple same lines; and print only the part of the string after the last slash (-F"/" use slash as delimiter; print $NF, print last field).
The output is gonna be written into file $FILE, which, if exists is gonna get overwritten.

Search for "sed regular expressions" for further explanations.

This script is incomplete, and won't work as is.