org.apache.drill.exec.physical.rowSet

Interface RowSetReader

All Superinterfaces:

ColumnReader, TupleReader

All Known Implementing Classes:

RowSetReaderImpl

public interface RowSetReader
extends TupleReader

Reader for all types of row sets: those with or without a selection vector. Iterates over "bare" row sets in row order. Iterates over selection-vector based row sets in selection vector order.

Method Summary

Modifier and Type	Method and Description
boolean	hasNext() Convenience method which whether the next call to next() will succeed.
int	hyperVectorIndex() Batch index: 0 for a single batch, batch for the current row is a hyper-batch.
int	logicalIndex() Gets the read position within the row set.
void	newBatch() Bind the reader to a new batch of data.
boolean	next() Advance to the next position.
int	offset() The index of the underlying row which may be indexed by an SV2 or SV4.
void	rewind() Reset the position to before the first row.
int	rowCount() Total number of rows in the row set.
void	setPosition(int index) Sets the iteration position.

Methods inherited from interface org.apache.drill.exec.vector.accessor.TupleReader

array, array, column, column, columnCount, dict, dict, scalar, scalar, tuple, tuple, tupleSchema, type, type, variant, variant

Methods inherited from interface org.apache.drill.exec.vector.accessor.ColumnReader

getAsString, getObject, isNull, schema, type

Method Detail

rowCount

int rowCount()

Total number of rows in the row set.

Returns:

total number of rows

hasNext

boolean hasNext()

Convenience method which whether the next call to next() will succeed. Purely optional.

Returns:

true if there is another record to read, false if not

next

boolean next()

Advance to the next position. If the underlying row set has a selection vector, then moves one position in the selection vector, and to whichever data record is indexed.

Returns:

true if another row is available, false if all rows have been read

logicalIndex

int logicalIndex()

Gets the read position within the row set. If the row set has a selection vector, this is the position in that vector; the actual record location will likely differ. Use offset() to get the actual row index.

Returns:

current iteration position

setPosition

void setPosition(int index)

Sets the iteration position. If the row set has a selection vector, this sets the index within that vector. The index must be from -1 to the rowCount() - 1. Set the value to one less than the position to be read in the next call to next(). An index of -1 means before the first row.

Parameters:

index - the desired index position

rewind

void rewind()

Reset the position to before the first row. Convenient method which is the same as setPosition(-1).

hyperVectorIndex

int hyperVectorIndex()

Batch index: 0 for a single batch, batch for the current row is a hyper-batch.

Returns:

index of the batch for the current row

offset

int offset()

The index of the underlying row which may be indexed by an SV2 or SV4.

Returns:

index of the underlying row

newBatch

void newBatch()

Bind the reader to a new batch of data. The vectors are unchanged, but the buffers are different. Assumes the schema has not changed: the columns and selection vector mode remain unchanged; only the buffers changed. If the schema changes, discard this reader and rebuild a new one.