--- Huge log files & DB search / Eugenio Gatto. - Torino, 1994-09-06
 
--> Nathan Brindle & LSTOWN-L
 
Without redefining log files as WEEKLY, you may use a search in the list
database to cut huge logs in smaller chronological portions, with
something like this:
 
//db JOB ECHO=NO
DATABASE SEARCH DD=cmd
//cmd DD *
SEARCH * IN listname SINCE yy-mm-dd UNTIL yy-mm-dd
INDEX
PRINT ALL OUTLIM=5000
/*
 
5000 lines (OUTLIM) roughly corresponds to 400 k: I don't know (can
anybody be more precise ?) if this OUTLIM may override smaller limits
set at higher levels. Anyway, should the output be truncated, it is easy
to choose a new starting point for next search from the list given by
INDEX (or simply ask for missing documents, with a subsequent PRINT
xxxx-yyyy, since item numbers are fixed, until back log files are not
deleted or modified).
 
Since notebook files automatically constitute a database, I suppose (but
I'd like to see it confirmed) that the procedure is efficient enough to
be encouraged as a good or better alternative to simple GET of full
logs. Of course, the technique for searching has to be teached to users,
but it is so more powerful than simple GET ...
 
<-- Eugenio Gatto    (CIDEM, Politecnico di Torino, Italia)
    TEL 39.11.564'6600 FACS 39.11.564'6609 EM [log in to unmask]