How to compare input lines but ignore decimals

lili2002 · 06-14-2013, 09:06 AM

If I have below lines in a file and I want to sort first and count unique ones but ignore decimals when comparing, what should i do in uniq options?

Error in test. start=5, executing=12, end=60
Error in test. start=1, executing=20, end=80
Error in test. start=10, executing=5, end=100
Warn in test. begin=1, end=2
Warn in test. begin=4, end=9

I want the output will be like:

3 Error in test. start=5, executing=12, end=60
2 Warn in test. begin=1, end=2

thedaver · 06-14-2013, 10:32 AM

Looks like you are counting matches of string "Error in test" and "Warn in test", right?

If you assume that both phrases end with the period "." character and you only want to count the instances, then this should work.

Code:

cat sourcefile.txt | grep -E "^Error in test|^Warn in test" | sort | uniq -c

allend · 06-14-2013, 10:35 AM

If you are only interested in the number of unique messages then

Code:

cat junk2.txt | cut -f1 -d '.' | uniq -c

where junk2.txt contains the input data you posted produces

Quote:

3 Error in test
2 Warn in test