Shell Programming and Scripting

I have a huge file and every paragraph has a date. But I want to retrieve the paras for the last two days only. So I can grep and findout the linenum for the first line since yesterday. Now I want to display everything after that line. And I am trying to do this inside a script so the linenum is a variable.
How can I do this. Or is there a way in GREP to find the first match for the pattern and then display the rest of the lines. This script is going to be for a heterogenous environment so if the solution is not applicable to a particular OS then I cannot use it.

Would it be possible to post some samples ?

not sure how your timestamp is formatted but you can:

grep -n timestamp file| head -1 | awk '{FS=":"; system("more +"$1" file|cat")}'

edit:
hmm i tried this on a test file, it's definately putting out the correct more statement but not displaying the file

grep -n "timestamp" file|head -1|awk '{FS=":"; print "more +" $1 " file|cat"}'

will output the correct command to run, i'm not sure why awk isn't outputting it when i run it through system.

Each para looks like this but with a diff date

2008-06-20-19.23.20.XXXXX-360 XXXXXXXXX LEVEL: Event
PID : 24009 TID : XXXXXXXX PROC : db2acd
INSTANCE: db2inst1 NODE : 000
FUNCTION: DB2 UDB, Health Monitor, db2HmonEvalBackup, probe:270
STOP : Automatic Backup: evaluation has finished with errors on database HASAMP

So the next para would be
2008-06-21-19.23.20.XXXXX-360 XXXXXXXXX LEVEL: Event
PID : 24009 TID : XXXXXXXXX PROC : db2acd
INSTANCE: db2inst1 NODE : 000
FUNCTION: DB2 UDB, Health Monitor, db2HmonEvalBackup, probe:270
STOP : Automatic Backup: evaluation has finished with errors on database HASAMP

I achived this by doing the following. Here $lfile refers to the file that I need to parse.

cat ${lfile} | wc -l | read totallines
cat ${lfile} | grep -m1 -n `date --date='1 day ago' +%F` | cut -d ":" -f1 | read linenum
lastlines=`expr $totallines - $linenum + 1`
cat ${lfile} | tail -$lastlines >> ${datafile}

I want to know if this can be done in one command. Because if the date doesnt exists then this will give me the entire file. So I would rather get and empty file than the whole file which tends to be very big.

This should do the job if your source file is a log file.

Code:

awk -v v=$(date --date='1 day ago' +%F) '$1 ~ v {f=1} f {print}' file > newfile