Shell Programming and Scripting

Hi,

I have a script that calls an external program to perform some calculations and then I read with "grep" and "sed" values from the output files. I've noticed that performance of KSH93 degrades with every iteration. The output files are all the same size, so I don't understand why after the first calculation it takes 5 seconds to read the values but after the 50th calculation it takes ~1 minute. Besides that, everything is working correctly, no error messages.

I had the same problem a couple of months ago with a simpler script. Back then I just executed the script with KSH88 and the problem was gone (I wasn't using any special features of KSH93).

But this actual script has gotten very complex using lots of KSH93 specific stuff, so I cannot use KSH88 here.

I've searched the internet concerning performance degradation, but found only a hint to a bug with the "unset" command. Well, I was using the "unset" command for some array variables, but even after commenting these lines out, the problem was still there.

Has anyone heard anything about KSH93 getting slower and slower with time?

By the way, I'm on a AIX 5.2 machine.

Thanks, Ingo

I have two output files. First I get some values from the first file, then I calculate values from the second file that are based on the first one. If you're wondering why I have semicolons at the end of each line, it's because the code is being joined in one string and executed by the "eval" command.

This code takes a couple of seconds for the first calculation and after the 50th it takes more than 1 minute, even though the output files are similar to each other in length.

# Timestamp
u_start=`perl -e 'print time()'`;

# Read values from EnvOPT.dat:? file
if [[ -a EnvOPT.dat:$OFP ]]; then
for x in `grep -e "^Contrail_Length " -e "^Contrail_Climate_Impact " -e "^Contrail_ETS_Cost " -e "^CO2_Emissions " -e "^CO2_Climate_Impact " -e "^CO2_ETS_Cost " EnvOPT.dat:$OFP`; do
param=`echo $x | awk '{print $1}'`;
value=`echo $x | awk '{print $3}'`;

case $param in
Contrail_Length) contrail_length="$value"; ;;
Contrail_Climate_Impact) contrail_climate_impact="$value"; ;;
Contrail_ETS_Cost) contrail_ets_cost="$value"; ;;
CO2_Emissions) co2_emissions="$value"; ;;
CO2_Climate_Impact) co2_climate_impact="$value"; ;;
CO2_ETS_Cost) co2_ets_cost="$value"; ;;
esac;
done;

# Edit both files, keep only important lines
sed '/#OFP_DEP_DEST#/,/ROUTING_TEXT/!d' ProfOPT.dat:$OFP | sed -e '1,12d' -e '$d' > ProfOPT.dat:tmp;
sed '/#Section MetPhenomIndices#/,$!d' EnvOPT.dat:$OFP | sed '1,3d' > EnvOPT.dat:tmp;

# Save edited files in array
set -A u_prof $(< ProfOPT.dat:tmp);
set -A u_env $(< EnvOPT.dat:tmp);

v=0;
# Number of lines (equal in both files)
u=${#u_prof[@]};
fuel_night=0;
while (( $v < $((u-1)) )); do
# Get values from first file (column 3)
val1=`echo ${u_env[$v]} | awk '{print $3}'`;
val2=`echo ${u_env[$((v+1))]} | awk '{print $3}'`;
# If values at line "i" and "i+1" equals "0" then...
if (( $val1 == 0 && $val2 == 0 )); then
# Get values from second file (column 13)
val1=`echo ${u_prof[$v]} | awk '{print $13}'`;
val2=`echo ${u_prof[$((v+1))]} | awk '{print $13}'`;
# Calculate delta fuel and add to total fuel
fuel_night=$((fuel_night+val2-val1));
fi;
((v+=1));
done;
fi;

u_end=`perl -e 'print time()'`;
echo " Elapsed time: $((u_end-u_start)) seconds";