Re: sort email by date

On 2/24/2008 9:47 PM, droid wrote:
> I have "received" emails from the early nineties to present.  These
> are concatenated together in two huge files.
>
> My problem is that segments of one file need to be inserted into the
> other file.  Also, there would be many dupes.
>
> I was hoping to find a utility that could do the sorting after joining
> them with 'cat', then use Thunderbird to remove
> the dupes.

How do you identify a dupe - by timestamp or by contents?

> But from the replies here, it appears this is more complicated than I
> supposed.

I misunderstood your original question. I think what you might want to do
is
split your 2 huge files into 1 file per email, named based on the
date+time of
the email plus some unique incrementing key. Then a simple "ls" will
"sort" them
and you can use "diff" as you create them (or afterwards) to decide if a
given
email with the same date/time as an existing one is actually a dup (in
which
case don't create the file) or a different email with the same timestamp
(in
which case increment the uniqueness key). Then you can "cat" them all
together
if you like and they'll automatically be sorted in the resulting file.

You may want to use procmail and formail to help normalise the emails for
selecting the timestamp and creating the unique files.

None of that necessarily requires awk, though, so further discussion of
that
solution would be OT for this NG. Follow up at comp.unix.shell if you want
to
pursue it.

	Ed.