I have the following script that is suppose to list files with a certain date string contained in the name of the file. The script won't list any of the files, even though they are there.

typeset -Z2 KEEPDAYS=$((`date +%d` -7))
MONTH=`date +%m`
YEAR=`date +%Y`

if [ $KEEPDAYS -le "0" ] ; then
MONTH=$((MONTH-1))
if [ $MONTH -eq "0" ] ; then
MONTH=12
YEAR=$((YEAR-1))
fi
set 'cal $MONTH $YEAR'
shift $(($# - 1))
KEEPDAYS=$1
fi

DATE=$YEAR$MONTH$KEEPDAYS

dir=$PWD
fileExp=$DATE

if [ "X${fileExp}" != "X" ] ; then
for file in `find . -name \*$fileExp\*`
do
echo $file
done
fi

I've also noticed that if I change the month and year to a leap year, say march 3, 2004. the value in KEEPDAYS comes back with garbage.

You don't say which Unix and shell you're using - looks like ksh or bash, but your script works fine on my Solaris 9 box using ksh. I suspect your problem is the way you are building the find command. I'd try replacing this:

if [ "X${fileExp}" != "X" ] ; then
for file in `find . -name \*$fileExp\*`
do
echo $file
done
fi

with this:

if [ ! -z ${fileExp} ]
then
find . -name "*${fileExp}*" -print|while read myfile
do
echo "${myfile}"
done
fi
# end stub

Regarding your second question, I don't understand. If you are going to use a date other than the system date, you have to find another method for determining KEEPDAYS other than the date command.

I'm sorry I just realized I stated my problem wrong. I need to list files that have a date that is older than DATE=$YEAR$MONTH$KEEPDAYS

I am using ksh on Unix.

Now, I'm confused how KEEPDAYS is determined.

Have you considered using the find command's modification time option:

-mtime n
True if the file's data was modified n days ago.

Also, there's a method for finding files modified within a given date range using find's -newer option. It's discussed here:

http://www.unixreview.com/documents...

KEEPDAYS is determined by the calender day - 7
typeset -Z2 KEEPDAYS=$((`date +%d` -7))

I then want to be able to list files that are older than the KEEPDAYS number.

Filenames are stored as PDFBOLS20061214165246310048A1DE00002. After the date and time stamp, the rest is randomly generated.
I need to be able to list any file that is older than the date the has been inserted into the filename '20061214'. So I should see '*20061207*' '*20061206*' '*20061205*' etc.

Using find -mtime would delete other files that should not be deleted. In the script before I could just delete anything just by using the just the year. I now need to keep the selected files if they are 0-7 days old. >7 They need to be listed and removed.

Why don't you use a combination of file name and modification time? Won't the modification time be clase to the time stamp embedded in the file name?

YEAR=2006
MONTH=12

FileExp=

find . -type f -name "PDFBOLS*${YEAR}${MONTH}*" -mtime ....

Thanks a lot for the help nails. I ended up using this:
YEAR=`date +%Y`

find $dir -type f -name "PDFBOLS${YEAR}*" ! -mtime -7 -print|sort|while read myfile
do
ls -l $myfile
rm -f $myfile
done

Thanks again.

you are welcome