Log analysis script in shell - shell

Question

I'm a newbie in scripting and I need your help. I have a log file, that I'm cleaned out. Looks like this (Time, duration(in millisec), action):
2012-04-28 00:00:00;277.406;
2012-04-28 00:00:00;299.680;
2012-04-28 00:00:00;282.338;
2012-02-28 00:00:00;272.241;
I need to make a script that use the duration data and count the action.

Answer 1

First - you need to make it easier to parse the different fields. A simple way is to change a semicolon to a space using
tr ";" " " <logfile|awkscript
Second, you need to create a table of low and high values. I'm using an associative array whose index is the name of the column. I do this in the BEGIN section.
You need to count when a value is within the low and high values. I do this in the middle section.
In the END section, I print out the values. I use 2 similar printf format strings to make sure the headers and values line up nicely:
#!/usr/bin/awk -f
BEGIN {
low["<1ms"]=0;high["<1ms"]=1
low["1-10ms"]=1;high["1-10ms"]=10
low["10-100ms"]=10;high["10-100ms"]=100
low["100-500ms"]=100;high["100-500ms"]=500
low[">500ms"]=500;high[">500ms"]=1000000000
}
{
# Middle section - for each line
duration=$3
for (i in high) {
if ((duration > low[i]) && (duration <= high[i]) ) {
# printf("duration: %d, low: %s,high: %s\n", duration, low[i], high[i]);
total+=duration # total duration
bin[i]++ # store a count into different bins
count++ # total number of measurements
}
}
}
END {
average=total/count
FMT="%-10s %10s %10s %10s %10s %10s\n"
NFMT="%-10.3f %10s %10s %10s %10s %10s\n"
printf(FMT,"AVG", "<1ms", "1-10ms", "10-100ms", "100-500ms", "500+ms")
printf(NFMT,average, bin["<1ms"], bin["1-10ms"], bin["10-100ms"], bin["100-500m\
s"], bin["500+ms"])
}
When I run this with your data, I get
AVG <1ms 1-10ms 10-100ms 100-500ms 500+ms
282.916 4