The input file is a SAM or ISAM file. It contains database and subschema records, which are structured as follows:
Database records:
DB'BLANK'dbname'BLANK'confname'BLANK'processor-name['BLANK'password]
Subschema records:
SS'BLANK'subschema-name'BLANK'dbname
Every record must begin in a new line and may be up to 160 characters in length. Leading blanks are ignored by UDS-D.
If the dbname, confname, processor-name and subschema-name are to be initially treated as locked, they must be identified by a preceding minus sign. These databases and subschemas will then be accessible to the application program only when the lock has been removed with the DAL command &UNLOCK DISTRIBUTION.
Database and subschema records that begin with an asterisk (*) are treated as comments and are skipped when creating the distribution table. This enables the database administrator to create an input file with a clean layout.
UDS-D checks the syntax and ensures that names are assigned uniquely. Only the following assignments are allowed:
subschema-name to only one dbname
dbname to only one confname/processor-name
If the same name occurs again, UDS-D ignores the corresponding database or subschema record and issues a message.
In the case of a multi-DB configuration, many DB records will have the same entry for confname/processor-name. If a password is to be assigned to such a configuration, the password only needs to be specified with one of the associated DB records.
If different passwords for a configuration are specified in different DB records, the first password specified in the sequence of DB records applies. The other conflicting passwords specified with the DB records belonging to the same configuration are rejected with an appropriate message.
Database records
Database records specify the DB configuration to which a database belongs, possibly the password for that configuration, and the host on which the configuration is located.
The following naming conventions apply:
processor-name
The name of the host processor may be up to eight characters in length.
dbname
Database names in the distribution table must be unique throughout the network. See also table 5 in chapter "SDF syntax representation".
confname
The first eight characters of the configuration-name, which must be unique in the first seven characters. The eighth character must not be a “@”. Note that zeros at the end of confname are not considered significant characters, i.e. no distinction is made between confname ABC and confname ABC0.
password
The cross-configuration password can have a maximum length of eight characters.
Subschema records
Subschema records specify the database to which the subschema belongs.
The following naming conventions apply:
subschema-name
This name must be unique in the first six characters within the distribution table throughout the network. See also table 5 in chapter "SDF syntax representation".
Only user subschemas may be specified. The PRIVACY-AND-IQF subschema and COMPILER subschema are not permitted in the distribution table.