Real Time Information Director User Documentation
    RTID Metadata Language 
Hewlett-Packard Company  57  529618-002   
3.  Create a PartitionParameters object with parameters consistent with those you 
specified in the KEYS table: 
•  Assign to the keySpaceName parameter the value you put in the NAME 
column for the namespace. 
•  Assign to the startingPartition parameter the smallest value you put in the 
PARTITIONID column for any row in the namespace. 
•  Assign to the numberPartitions parameter the value you put in the 
MAX_PARTIONID column for the namespace. 
•  Assign to the partionRange parameter the value you put in the MAXKEY 
column for the namespace. 
4.  Create a PartitionedKeyDispenser object, specifying the keyspace name, the 
partitioning parameters, and the number of keys to be fetched in a block, for each 
partition. 
5.  Create an EnrichedRecord that refers to the PartitionedSystemKey enrichment 
class. Specify the partitioning parameters, the column that will contain the system 
key, the column that will contain the partitioning ID, the name of the key 
dispenser, and the partitioning strategy you’ve decided to use for the type of data. 
Note:  If you’re using a Mapping enrichment to correlate internal with external IDs, 
you’ll want to provide for the mapping table to be updated automatically when a new 
system key, corresponding to an internal ID, is assigned. You can achieve this goal by 
making defining the mapping table, in your content metadata, as a SplitRecord in relation 
to the record for which the PartitionedSystemKey enrichment is invoked. 
The contructors for the classes mentioned here are described in the RTID Javadocs. For 
examples, see “Partitioning Tables,” later in this section. 
InstanceId 
The InstanceID enrichment assigns a line number or subitem number to a specified 
column of a record that is logically a child to another record.  All the children of the same 
parent record in an insert or delsert document are automatically assigned numbers in 
ascending sequence starting at 1. For a document that is upserted, the previously inserted 
children of the same parent record are checked to determine the next available number to 
assign to new children. But if the number is specified in the input, the child is meant to 
update or replace the previous child with the same number. Conversely, omitting this 
field in the input, or setting it to zero, indicates that this is a new child. InstanceId is 
often used in conjunction with the SystemKey of the parent record to form the key fields 
of a record. When records are nested to many levels in a document, the key of a record 
can contain several instance IDs inherited from ancestors. 
The InstanceID constructor looks like this: 










