Hi!
I am new to shell scripting and am having some trouble implementing a simple file purge script. What I want to do is search for a word in a file ie a datestamp and then delete all lines in the file before that specific datestamp.

What I have so far is a grep -n command to get the line number/ It comes in the format 'lineumber:datestamp' is there anyway to get rid of the ':datestamp' portion.

Once I have this linenumber is there an sed command that could be used to delete all lines up to this point?
Any suggestions would be very helpful!
Thanks in Advance,
JK

You can branch within sed, but I believe an easier way to do this is to use combinations of head/tail to isolate what you want, and then rewrite the file.

re head/tail; yep, that'll work. Look up the cut cmd (man cut), specifically the -d option to split the value on the ':'

Thanks guys for the suggestions what I have done now is use a grep -n and a cut command to isolate the linenumber followed by a head command in case there is more than one instance of the pattern (I only want the first) and then pass this variable to an sed command and it works fine


grep -n "pattern" log.txt | cut -f1 -d:>TEST.TXT
firstOccurenceLineNum=`head -1 TEST.TXT`


# Get linenum-1 of first instance and delete up to there
firstOccurenceLineNum=`expr $firstline - 1`
sed "1,$firstline d" <log.txt >TEST.TXT