org.apache.drill.exec.physical.resultSet.impl

Class RowSetLoaderImpl

java.lang.Object

org.apache.drill.exec.vector.accessor.writer.AbstractTupleWriter

org.apache.drill.exec.physical.resultSet.impl.RowSetLoaderImpl

All Implemented Interfaces:

RowSetLoader, ColumnWriter, TupleWriter, WriterEvents, WriterPosition

public class RowSetLoaderImpl
extends AbstractTupleWriter
implements RowSetLoader

Implementation of the row set loader. Provides row-level operations, leaving the result set loader to provide batch-level operations. However, all control operations are actually delegated to the result set loader, which handles the details of working with overflow rows.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.drill.exec.vector.accessor.writer.AbstractTupleWriter

AbstractTupleWriter.TupleObjectWriter, AbstractTupleWriter.TupleWriterListener

Nested classes/interfaces inherited from interface org.apache.drill.exec.vector.accessor.TupleWriter

TupleWriter.UndefinedColumnException

Nested classes/interfaces inherited from interface org.apache.drill.exec.vector.accessor.writer.WriterEvents

WriterEvents.ColumnWriterListener, WriterEvents.State

Field Summary

Fields inherited from class org.apache.drill.exec.vector.accessor.writer.AbstractTupleWriter

childIndex, listener, logger, state, tupleSchema, vectorIndex, writers

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	RowSetLoaderImpl(ResultSetLoaderImpl rsLoader, TupleMetadata schema)

Method Summary

Modifier and Type	Method and Description
RowSetLoader	addRow(Object... values) Write a row of values, given by Java objects.
RowSetLoader	addSingleCol(Object value) Similar to RowSetLoader.addRow(Object...), but for the odd case in which a row consists of a single column that is an object array (such as for a list or map) and so is ambiguous.
void	endBatch()
boolean	isFull() Indicates that no more rows fit into the current row batch and that the row batch should be harvested and sent downstream.
ResultSetLoader	loader()
int	rowCount() The number of rows in the current row set.
int	rowIndex() The index of the current row.
void	save() Saves the current row and moves to the next row.
ColumnMetadata	schema() Returns the schema of the column associated with this writer.
boolean	start() Prepare a new row for writing.

Methods inherited from class org.apache.drill.exec.vector.accessor.writer.AbstractTupleWriter

addColumn, addColumn, addColumnWriter, array, array, bindIndex, bindIndex, bindListener, bindListener, column, column, copy, dict, dict, dump, endArrayValue, endWrite, isProjected, isProjected, lastWriteIndex, listener, nullable, postRollover, preRollover, restartRow, rowStartIndex, saveRow, scalar, scalar, set, setNull, setObject, size, startRow, startWrite, tuple, tuple, tupleSchema, type, type, type, variant, variant, writeIndex

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

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

addColumn, addColumn, array, array, column, column, dict, dict, isProjected, scalar, scalar, set, size, tuple, tuple, tupleSchema, type, type, variant, variant

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

copy, isProjected, nullable, setNull, setObject, type

Constructor Detail

RowSetLoaderImpl

protected RowSetLoaderImpl(ResultSetLoaderImpl rsLoader,
                           TupleMetadata schema)

Method Detail

loader

public ResultSetLoader loader()

Specified by:

loader in interface RowSetLoader

addRow

public RowSetLoader addRow(Object... values)

Description copied from interface: RowSetLoader

Write a row of values, given by Java objects. Object type must match expected column type. Stops writing, and returns false, if any value causes vector overflow. Value format:

• For scalars, the value as a suitable Java type (int or Integer, say, for INTEGER values.)
• For scalar arrays, an array of a suitable Java primitive type for scalars. For example, int[] for an INTEGER column.
• For a Map, an Object array with values encoded as above. (In fact, the list here is the same as the map format.
• For a list (repeated map, list of list), an Object array with values encoded as above. (So, for a repeated map, an outer Object map encodes the array, an inner one encodes the map members.

Specified by:

addRow in interface RowSetLoader

Parameters:

values - variable-length argument list of column values

Returns:

this writer

addSingleCol

public RowSetLoader addSingleCol(Object value)

Description copied from interface: RowSetLoader

Similar to RowSetLoader.addRow(Object...), but for the odd case in which a row consists of a single column that is an object array (such as for a list or map) and so is ambiguous.

Specified by:

addSingleCol in interface RowSetLoader

Parameters:

value - value of the one and only column

Returns:

this writer

rowIndex

public int rowIndex()

Description copied from interface: RowSetLoader

The index of the current row. Same as the row count except in an overflow row in which case the row index will revert to zero as soon as any vector overflows. Note: this means that the index can change between columns in a single row. Applications usually don't use this index directly; rely on the writers to write to the proper location.

Specified by:

rowIndex in interface RowSetLoader

Returns:

the current write index

save

public void save()

Description copied from interface: RowSetLoader

Saves the current row and moves to the next row. Failing to call this method effectively abandons the in-flight row; something that may be useful to recover from partially-written rows that turn out to contain errors. Done automatically if using setRow().

Specified by:

save in interface RowSetLoader

start

public boolean start()

Description copied from interface: RowSetLoader

Prepare a new row for writing. Call this before each row.

Handles a very special case: that of discarding the last row written. A reader can read a row into vectors, then "sniff" the row to check, for example, against a filter. If the row is not wanted, simply omit the call to save() and the next all to start() will discard the unsaved row.

Note that the vectors still contain values in the discarded position; just the various pointers are unset. If the batch ends before the discarded values are overwritten, the discarded values just exist at the end of the vector. Since vectors start with garbage contents, the discarded values are simply a different kind of garbage. But, if the client writes a new row, then the new row overwrites the discarded row. This works because we only change the tail part of a vector; never the internals.

Specified by:

start in interface RowSetLoader

Returns:

true if another row can be added, false if the batch is full

endBatch

public void endBatch()

isFull

public boolean isFull()

Description copied from interface: RowSetLoader

Indicates that no more rows fit into the current row batch and that the row batch should be harvested and sent downstream. Any overflow row is automatically saved for the next cycle. The value is undefined when a batch is not active.

Will be false on the first row, and all subsequent rows until either the maximum number of rows are written, or a vector overflows. After that, will return true. The method returns false as soon as any column writer overflows even in the middle of a row write. That is, this writer does not automatically handle overflow rows because that added complexity is seldom needed for tests.

Specified by:

isFull in interface RowSetLoader

Returns:

true if another row can be written, false if not

rowCount

public int rowCount()

Description copied from interface: RowSetLoader

The number of rows in the current row set. Does not count any overflow row saved for the next batch.

Specified by:

rowCount in interface RowSetLoader

Returns:

number of rows to be sent downstream

schema

public ColumnMetadata schema()

Description copied from interface: ColumnWriter

Returns the schema of the column associated with this writer.

Specified by:

schema in interface ColumnWriter

Returns:

schema for this writer's column