Rediger

Del via


Batch Processing (Analysis Services)

Applies to: SQL Server Analysis Services Azure Analysis Services Fabric/Power BI Premium

In SQL Server Analysis Services, you can use the Batch command to send multiple processing commands to the server in a single request. Batch processing gives you a way to control which objects are to be processed, and in what order. Also, a batch can run as a series of stand-alone jobs, or as a transaction in which the failure of one process causes a rollback of the complete batch.

Batch processing maximizes data availability by consolidating and reducing the amount of time taken to commit changes. When you fully process a dimension, any partition using that dimension is marked as unprocessed. As a result, cubes that contain the unprocessed partitions are unavailable for browsing. You can address this with a batch processing job by processing the dimensions together with the affected partitions. Running the batch processing job as a transaction makes sure that all objects included in the transaction remain available for queries until all processing is completed. As the transaction commits the changes, locks are put on the affected objects, making the objects temporarily unavailable, but overall the amount of time used to commit the changes is less than if you processed objects individually.

The procedures in this topic show the steps for fully processing dimensions and partitions. Batch processing can also include other processing options, such as incremental processing. For these procedures to work correctly, you should use an existing SQL Server Analysis Services database that contains at least two dimensions and one partition.

This topic includes the following sections:

Batch Processing in SQL Server Data Tools

Batch Processing using XMLA in Management Studio

Batch Processing in SQL Server Data Tools

Before objects can be processed in SQL Server Data Tools, the project that contains the objects must be deployed. For more information, see Deploy Analysis Services Projects (SSDT).

  1. Open SQL Server Data Tools.

  2. Open a project that has been deployed.

  3. In Solution Explorer, under the deployed project, expand the Dimensions folder.

  4. Holding the Ctrl key, click each dimension listed in the Dimensions folder.

  5. Right-click the selected dimensions, and then click Process.

  6. Holding the Ctrl key, click each dimension listed in the Object list.

  7. Right-click the selected dimensions and select Process Full.

  8. To customize the batch process job, click Change Settings.

  9. Under Processing options, mark the following settings:

    • Processing Order is set to Sequential, and Transaction mode is set to One Transaction.

    • Writeback Table Option is set to Use existing.

    • Under Affected Objects, select the Process affected objects check box.

  10. Click the Dimension key errors tab. Verify that Use default error configuration is selected.

  11. Click OK to close the Change Settings screen.

  12. Click Run in the Process Objects screen to start the processing job.

  13. When the Status box shows Process succeeded, click Close.

  14. Click Close on the Process Objects screen.

Batch Processing using XMLA in Management Studio

You can create an XMLA script that performs batch processing. Start by generating an XMLA script in Management Studio for each object, and then combine them into a single XMLA Query that you run interactively or inside a scheduled task.

For step by step instructions, see Example 2 in Schedule SSAS Administrative Tasks with SQL Server Agent

See Also

Processing a multidimensional model (Analysis Services)