Interface TableFunctionProcessorState

All Known Implementing Classes:
TableFunctionProcessorState.Blocked, TableFunctionProcessorState.Finished, TableFunctionProcessorState.Processed

The result of processing input by TableFunctionDataProcessor or TableFunctionSplitProcessor. It can optionally include a portion of output data in the form of Page The returned Page should consist of: - proper columns produced by the table function - one column of type BIGINT for each table function's input table having the pass-through property (see TableArgumentSpecification.isPassThroughColumns()), in order of the corresponding argument specifications. Entries in these columns are the indexes of input rows (from partition start) to be attached to output, or null to indicate that a row of nulls should be attached instead of an input row. The indexes are validated to be within the portion of the partition provided to the function so far. Note: when the input is empty, the only valid index value is null, because there are no input rows that could be attached to output. In such case, for performance reasons, the validation of indexes is skipped, and all pass-through columns are filled with nulls.