I suppose the file is already sorted implicitly, because messages are stored sequentially in a time-wise order. This makes your task easier, because you can simply search for the first entry of a given date and delete all lines from there (or up to this line, depending on the file being sorted ascendingly or descendingly).

Suppose your date to search for would be stored in "$ThresholdDate" your sed-command could look like:

Code:

sed '/^'"$ThresholdDate"'/q' /path/to/infile > /path/to/outfile

This would print all lines up to (but not including) the first line containing "$TresholdDate" at the line start and then stop.

Two problems: you have to calculate the threshold-date and store it to "$ThresholdDate" and second, you have to escape all the characters with special meanings to sed which you use in your variable, especially the slashes ("/").

To calculate the correct date store the current date in several variables for year, month and day, then subtract 30 days and reassemble that to the result. Depending on how exact your result has to be you could simplify this by ignoring odd years, make every month 30 days long, etc. In either case it will be simply a matter of some additions and subtractions.

You could, of course, also use "datecalc" written by our venerable brother perderabo.

To escape the slashes add "\/" instead of the regular "/" into the reassembled date:

Code:

ThresholdDate="${month}\/${day}\/${year}"

I hope this helps.

bakunin