Download LCD User`s Guide

Transcript
•
commit updates to an Archive Manager database
A full description of the new application program interface may be found
in chapter 4 of this manual.
Batch processing
An API is supplied to enable batch programs to write objects to any
storage level in an Archive Manager database. The batch program will
identify the archive object by primary key and archive date. One or more
secondary keys for each object may additionally be identified
automatically or explicitly. This information is used by Archive Manager
to maintain primary and secondary indexes for the database. These
indexes are used to enable rapid retrieval of data in response to a batch
or online object retrieval request.
Creation and maintenance of all components of an Archive Manager
database is performed by the product itself. No direct user access is
required to any Archive Manager system dataset.
The interface allows database updates to be explicitly committed at any
point. Updates may also be implicitly committed when a database is
closed for output processing. A restart facility is provided in order to
facilitate rerun or restart of batch archival procedures after job failure,
without the requirement of any prior data recovery operation.
The Archive Manager API can be invoked from a batch program to
retrieve objects from an Archive Manager database. Objects to be
retrieved are identified by primary key and archive date.
Each logical record within an identified object is returned sequentially to
the calling application in response to a program call. An indication is
given by Archive Manager when all logical records for the specified object
have been retrieved. Direct or sequential retrieval of objects may be
performed using the Archive Manager batch retrieval interface.
Access to all components of an Archive Manager database, including
dataset allocation and de-allocation, and data retrieval are supplied by
Archive Manager. No direct user access is required to any Archive
Manager system dataset.
Batch retrieval in this manner may be used for producing printed reports
on archived data, re-printing archived bills etc. This batch process might
be requested from an online application by users wishing to recall multiple
Archive Manager objects for a specific primary key, and who did not wish
to wait for synchronous (online) retrieval to display all the required details
on the user's terminal. The asynchronous (batch) retrieval could run as a
background batch task, printing the requested details on a local printer
while the online user continues to use the online system.
Archive Manager User Guide
15