|
migrate!DATA
MSS data migration software
provides and automated solution to the problem of defining and populating
a relational database from a mainframe, hierarchical (DMSII)
database.
The target database systems
addressed natively are Oracle, SQL/Server, and IBM's DB2. Others can be handled using common
interchange formats, e.g. csv files.
migrate!DATA key features
include:
·
Creation
of an SQL database definition for the new relational database. migrate!DATA will produce an
equivalent to the DMSII data even handling variable format records,
groups, occurs and embedded structures.
·
Automation
of the load process by creation of control files and a script that will
run the standard load utilities for Oracle, DB2 or SQL/Server.
·
Data
translation to a valid, unambiguous form that can easily be transported
to an open platform. We use
display data with separate signs for numeric items and translate to
ascii but take account of special language issues. migrate!DATA handles
Arabic, Turkish, Hebrew and other character sets.
·
The
process is highly automated since it may be needed often. During system testing of a migration
many database dumps, reflecting the data state at different points in a
test sequence, may be required.
·
The
intermediate data medium is designed to be conveniently transportable
and multiple transport mechanisms are provided. An entire database or a selection of
structures may be dumped to a single flat disk file and then sent, e.g.
via ftp, to the target machine. Or data can be extracted directly to
tape. Or a socket connection to a remote server program (provided) can
be used.
·
Data
can be dumped subject to complex selection and individual records
included or excluded according to user supplied criteria. Individual fields can also be dropped
from the DMSII data if required. The files to be dumped will be
specified via parameters and data items may be excluded selectively.
·
The
order in which data records are extracted is specifiable in different
ways: by physical order, via an index structure or by a generated sort.
·
migrate!DATA
is also aware of some LINC-specific issues. For instance, on
recognising the Linc name given to the link to a memo data structure,
will generate code to merge the memo data with the rest of the
dataset’s data in the relational table row. Appropriate ddl definitions are
created to make the table one physical structure instead of two (of
course, the program always ‘thought’ it was one structure anyway).
|
