Techniques for managing XML data associated with multiple execution units

A non-transitory computer-readable storage medium storing one or more sequences of instructions which, when executed by one or more processors, cause performance of steps comprising:

• analyzing a query execution plan that specifies tasks to be performed by a plurality of execution units;
• wherein the plurality of execution units include a first execution unit and a second execution unit;
• wherein the query execution plan is to service a request for data;
• wherein analyzing the query execution plan includes:
• detecting that a portion of the query execution plan specifies first tasks to be performed by the first execution unit;
• detecting that another portion of the query execution plan specifies second tasks to be performed by the second execution unit;
• and
• detecting that, as a result of performing the first tasks, the first execution unit will generate XML data the second execution unit needs to perform the second tasks;
• generating information to send to said first execution unit to cause said first execution unit to perform first tasks;
• wherein in the absence of any annotation, said information would cause said first execution unit to generate the XML data in a first form that cannot be used by said second execution unit;
• wherein, in said first form, the XML data includes one or more references to execution unit-specific data that is accessible by the first execution unit but that is not accessible by the second execution unit;
• and
• annotating said information with an annotation that causes the XML data generated by said first execution unit to be transformed to a second form for use by said second execution unit in performing said second tasks;
• wherein, in said second form, the XML data does not include any references to execution unit-specific data.

A method for processing XML data, comprising:

• analyzing a query execution plan that specifies tasks to be performed by a plurality of execution units;
• wherein the plurality of execution units include a first execution unit and a second execution unit;
• wherein the query execution plan is to service a request for data;
• wherein analyzing the query execution plan includes:
• receiving information at the first execution unit to cause said first execution unit to perform first tasks;
• detecting that another portion of the query execution plan specifies second tasks to be performed by the second execution unit;
• and
• detecting that, as a result of performing the first tasks, the first execution unit will generate XML data that the second execution unit needs to perform the second tasks;
• wherein, in the absence of any annotation, said information would cause said first execution unit to generate the XML data in a first form that cannot be used by said second execution unit;
• wherein, in said first form, the XML data includes one or more references to execution unit-specific data that is accessible by the first execution unit but that is not accessible by the second execution unit;
• annotating said information with an annotation that causes the XML data generated by said first execution unit to be transformed to a second form for use by said second execution unit in performing said second tasks;
• wherein, in said second form, the XML data does not include any references to execution unit-specific data;
• and
• providing XML data that is annotated to said second execution unit in said second form for use in performing said second tasks by said second execution unit;
• wherein the method is performed by one or more computing devices.

A non-transitory computer-readable storage medium storing one or more sequences of instructions which, when executed by one or more processors, cause performance of steps comprising:

• analyzing a query execution plan that specifies tasks to be performed by a plurality of execution units;
• wherein the plurality of execution units include a first execution unit and a second execution unit;
• wherein the query execution plan is to service a request for data;
• wherein analyzing the query execution plan includes:
• receiving information at the first execution unit to cause said first execution unit to perform first tasks;
• detecting that another portion of the query execution plan specifies second tasks to be performed by the second execution unit;
• and
• detecting that, as a result of performing the first tasks, the first execution unit will generate XML data that the second execution unit needs to perform the second tasks;
• wherein in the absence of any annotation, said information would cause said first execution unit to generate the XML data in a first form that cannot be used by said second execution unit;
• wherein, in said first form, the XML data includes one or more references to execution unit-specific data that is accessible by the first execution unit but that is not accessible by the second execution unit;
• annotating said information with an annotation that causes the XML data generated by said first execution unit to be transformed to a second form for use by said second execution unit in performing said second task;
• wherein, in said second form, the XML data does not include any references to execution unit-specific data;
• and
• providing XML data that is annotated to said second execution unit in said second form for use in performing said second tasks by said second execution unit.