Shell Programming and Scripting

This has probably been asked 100 times, but I couldn't find any articles on point. I have a script that runs on the last day of every month at 11:30pm. If cats a number of input tables that were created the previous month (or earlier), combines them into one master file and erases the indivual files. It also compresses any master files from previous months. My problem is that if the script is run on August 31, it will take all the files from July, and make a master file with the name EKVI0804.DAT

What I want to do is a have it use the previous month (EKVI0704.DAT) for the file (since the files belong to July). Below is my current script. Any help would be appreciated.

==========================
#!/bin/ksh

# Because a cron job can not be told to run on the last day of a month,
# it will be run on the 28-31 of every month at 11:30pm. The following
# code will determine if we are on the last day of the month. Because a
# cron job doesn't have any environmental variables available from the
# submitting session, all paths need to be hard coded. If this job needs
# to be run for a gers production except live, please make the
# approporate modifications to the path names. The test will be for any
# file that was created before the current month.

days_in_month=$(echo $(cal) | awk '{print $NF}')
day_of_month=$(date +%d)
if [[ $days_in_month -eq $day_of_month ]]
then
touch -m -t `date +%Y%m`010000.01 /gers/live/adhoc/ekvidatetestfile
cd /gers/live/datafiles
find . \( ! -name . -prune \) -name "EKVI*\.DAT" -type f ! -newer /gers/live/adhoc/ekvidatetestfile -exec compress {} \;
cd /gers/live/adhoc
find . \( ! -name . -prune \) -name "EKVI*" -type f ! -newer /gers/live/adhoc/ekvidatetestfile -exec cat {} > /gers/live/datafiles/EKVI`date +%m%y`.DAT \;
find . \( ! -name . -prune \) -name "EKVI*" -type f ! -newer /gers/live/adhoc/ekvidatetestfile -exec rm {} \;
rm /gers/live/adhoc/ekvidatetestfile
else
echo not the last day of the month
exit
fi

I figured out how to do it. It might not be eligent, but it works.

===============================
#!/bin/ksh

# Because a cron job can not be told to run on the last day of a month,
# it will be run on the 28-31 of every month at 11:30pm. The following
# code will determine if we are on the last day of the month. Because a
# cron job doesn't have any environmental variables available from the
# submitting session, all paths need to be hard coded. If this job needs
# to be run for a gers production except live, please make the
# approporate modifications to the patch names. The test will be for any
# file that was created before the current month.

days_in_month=$(echo $(cal) | awk '{print $NF}')
day_of_month=$(date +%d)

last_month=$(expr `date +%m` - 1)
last_year=$(date +%y)
if [[ $last_month -eq 0 ]]
then
last_month=12
last_year=$(expr $last_year - 1)
fi
if [[ $last_month -lt 10 ]]
then
last_month="0"$last_month
fi

file_name=$last_month$last_year

if [[ $days_in_month -eq $day_of_month ]]
then
touch -m -t `date +%Y%m`010001.01 /gers/live/adhoc/ekvidatetestfile
cd /gers/live/datafiles
find . \( ! -name . -prune \) -name "EKVI*\.DAT" -type f ! -newer /gers/live/adhoc/ekvidatetestfile -exec compress {} \;
cd /gers/live/adhoc
find . \( ! -name . -prune \) -name "EKVI*" -type f ! -newer /gers/live/adhoc/ekvidatetestfile -exec cat {} > /gers/live/datafiles/EKVI$file_name.DAT \;
find . \( ! -name . -prune \) -name "EKVI*" -type f ! -newer /gers/live/adhoc/ekvidatetestfile -exec rm {} \;
rm /gers/live/adhoc/ekvidatetestfile
else
echo not the last day of the month
exit
fi