DTS Parallel Read Maintenance
This chapter describes the screen for maintaining the parallelization of data selection tasks. To define the parallelization, see the chapter Setting up the Parallel Read. DTS contains three task types that can be parallelized:
-
Preselection tasks. Can be maintained in the step Maintain derivation paths.
-
Read task parallelization. Can be maintained in the step Maintain technical settings.
-
Delta read tasks. Can be maintained in the step Maintain technical settings.
The pop-up with the parallelization setup is displayed by clicking the button in the respective maintenance screen.
The following options are available:
-
Partitioned table: Non-editable field. Specifies the name of the table for which the partitioning is being set up.
-
Shadow table: Available for the trigger-based NZD setup only. Specifies the database trigger shadow table related to the table.
-
Partitioning active: Flag indicating whether the partitioning is used for the generation of tasks and programs. If the parallel read has been disabled for a table, you must regenerate the read task (via the step Generate transformation tasks) and the read program (via the step Generate programs) for the table.
NOTE Any already built partitioning will be deleted if saved with this setting. You can close the screen by clicking Cancel or build it again via the step Execute parallel read builder. -
Active delta run partitioning type: Available for the parallelization of delta read tasks only. Specifies the actual delta to which the partitioning will be applied.
-
Number of generated partitions: Specifies the number of generated partitions. Visible only if the partitions have been built. In other cases, this parameter is hidden.
-
Partitioning type: Available for the parallelization of read tasks only.
-
Build partitioning: Sets up the partitioning for this table.
-
Reuse partitioning: Enables you to copy the partitioning from another table. A parameter to enter the source table is displayed.
-
-
Partitioning options: Available for the partitioning type Build partitioning only.
-
Partitioning fields: Lists of fields to be used for the partitioning. The best practice is to use all key fields. For more information, see the button Set all key fields below.
WARNING If you select fields other than key fields, it might result in worse overall performance. Therefore, you need a good understanding of database processing principles in order to make the selection. Typically, a secondary index must exist for the table. -
Number of partitions: Allows you to enter an integer value. The table will be split into the entered number of partitions. The default value is 10.
-
Rows per partition: Allows you to enter an integer value. The partitioning build process will start a new partition after the number of records defined in the parameter is reached.If both parameters are filled, the option Rows per partition is considered.
-
Use preselection: For a detailed description of this option, see the step Analyze technical settings of tables.
-
-
Set all key fields (Shift+F1): Populates the partitioning fields select option with the key fields of the table. We recommend using this option. The DTS partition builder splits the table by fields automatically until it reaches the partitioning condition defined by the user. For example, if the first field has enough cardinality, only the first field is used. If not, the second field is taken into consideration and so on until the defined conditions are met.
-
View (F8): Shows the build partitions for the table.
-
Delete (Shift+F2): Deletes the partitioning setup.
-
Save (Enter): Saves the setup.