Red Hat Training

A Red Hat training course is available for Red Hat JBoss Data Virtualization

3.12. Partial Results Mode

JBoss Data Virtualization supports a partial results query mode. This mode changes the behavior of the query processor so the server returns results even when some data sources are unavailable.
For example, suppose that two data sources exist for different suppliers and your data designers have created a virtual group that creates a union between the information from the two suppliers. If your application submits a query without using partial results query mode and one of the suppliers'€™ databases is down, the query against the virtual group returns an exception. However, if your application runs the same query in partial results query mode, the server returns data from the running data source and no data from the data source that is down.
When using partial results mode, if a source throws an exception during processing it does not cause the user's query to fail. Rather, that source is treated as returning no more rows after the failure point. Most commonly, that source will return 0 rows.
This behavior is most useful when using UNION or OUTER JOIN queries as these operations handle missing information in a useful way. Most other kinds of queries will simply return 0 rows to the user when used in partial results mode and the source is unavailable.

Note

In some instances, (typically when you are using JDBC sources), if the source is not available initially, its absence will prevent Teiid from automatically determining the appropriate set of source capabilities. If you see an exception indicating that the capabilities for an unavailable source are not valid in partial results mode, then it may be necessary to manually set the database version or similar property on the translator to ensure that the capabilities are known even if the source is not available.