This file is C$DOC:CDFSGA_P_M_DATABASES.NOTE


    Copies of the CDF Production Manager Databases are
maintained on cdfsga.   These databases tell which files
appear on which tapes.   A common use of these databases 
is to generate UICs for jobs in automatic ways.

    Please look in directory
~cdfman/p_m/production/database
for your favorite data stream.

    We have found that it works better in practice when
one specifies both the tape and the files desired from
those tapes (rather than using tape wildcards.)

    The reason is that sometimes tapes have bad files on them.
These bad files are removed from  the Production Manager
Databases, but are seen when tape wildcards are used, leading
to duplicate events and partial files.

    Many people have found it convenient to use the
awk programming language to parse through the
production manager databases, and have found it
convenient to split the records into fields 
by taking the | character as the field seperator.  The
qualfiers for such use are:
    awk -F/|