Hi Paul,

On Tue 2036-Dec-16 12:03, Paul Quinn (3:640/384) wrote to Richard Webb:

 RW> THe objective:  text Files are combined which contain data
 RW> entries.  Every 12 hours the file needs to be rebuild from
 RW> files containing data received or entered during the last 48 hours.
 RW> Which means, we kill files older than 48 hours, rebuild.

 RW> AS earlier noted, this task would run twice a day.
PQ> Which part of the process works already (scheduling, rebuild -or-
PQ> kill), Richard?  :-)

scheduling of course works.

At midnight and noon gmt scheduling in place for other tasks as well.  Data
will be coming in periodically for these text files, and we want to purge
the oldest.

PQ> Killing off old files is relatively easy.  I do something similar in
PQ> my toss batch for the Fidonews file echo:

YEp, I"m cogitating on a couple of approaches now, not quite sure how
I want to go about it.
NEW data would be written to today's.  nExt event, a new
"current" file is the one appended.


PQ> Mind you, it's made a little easier by knowing the _number_ of files
PQ> involved.  This method of file/filename manipulation with DIR & NSET
PQ> could also be used for the rebuild as well, I guess, if a rebuild is
PQ> done whenever a new (single) data file is received/generated.
close.  WE'd have, every 12 hours a new current file being
appended whenever data received.

DAta in this case is information on vessels offshore.
The system receives the information, when found the batch
invoked uses sed to grab all data and none of the header
stuff to the current file.

NExt scheduled event, new current file, of course.  Purge
the file with data over 48 hours old.

WHen another user requests this info by email, or I want to
have a look at it, all files are combined into one text
file.


PQ> Is this close to what you're after?

SEems to be.  Any thought after a bit more description?

Regards,
           Richard
--- timEd 1.10.y2k+