org.apache.orc.impl.writer

Interface TreeWriter

All Known Implementing Classes:

BinaryTreeWriter, BooleanTreeWriter, ByteTreeWriter, CharTreeWriter, DateTreeWriter, Decimal64TreeWriter, DecimalTreeWriter, DoubleTreeWriter, FloatTreeWriter, IntegerTreeWriter, ListTreeWriter, MapTreeWriter, StringBaseTreeWriter, StringTreeWriter, StructTreeWriter, TimestampTreeWriter, TreeWriterBase, UnionTreeWriter, VarcharTreeWriter

public interface TreeWriter

The writers for the specific writers of each type. This provides the generic API that they must all implement.

Nested Class Summary

Nested Classes

Modifier and Type	Interface and Description
static class	TreeWriter.Factory

Method Summary

Modifier and Type	Method and Description
void	createRowIndexEntry() Create a row index entry at the current point in the stripe.
long	estimateMemory() Estimate the memory currently used to buffer the stripe.
void	flushStreams() Flush the TreeWriter stream
long	getRawDataSize() Estimate the memory used if the file was read into Hive's Writable types.
void	updateFileStatistics(OrcProto.StripeStatistics stripeStatistics) During a stripe append, we need to update the file statistics.
void	writeBatch(ColumnVector vector, int offset, int length) Write a ColumnVector to the file.
void	writeFileStatistics(OrcProto.Footer.Builder footer) Add the file statistics to the file footer.
void	writeRootBatch(VectorizedRowBatch batch, int offset, int length) Write a VectorizedRowBath to the file.
void	writeStripe(OrcProto.StripeFooter.Builder stripeFooter, OrcProto.StripeStatistics.Builder stats, int requiredIndexEntries) Write the stripe out to the file.

Method Detail

estimateMemory

long estimateMemory()

Estimate the memory currently used to buffer the stripe.

Returns:

the number of bytes

getRawDataSize

long getRawDataSize()

Estimate the memory used if the file was read into Hive's Writable types. This is used as an estimate for the query optimizer.

Returns:

the number of bytes

writeRootBatch

void writeRootBatch(VectorizedRowBatch batch,
                    int offset,
                    int length)
             throws IOException

Write a VectorizedRowBath to the file. This is called by the WriterImplV2 at the top level.

Parameters:

batch - the list of all of the columns

offset - the first row from the batch to write

length - the number of rows to write

Throws:

IOException

writeBatch

void writeBatch(ColumnVector vector,
                int offset,
                int length)
         throws IOException

Write a ColumnVector to the file. This is called recursively by writeRootBatch.

Parameters:

vector - the data to write

offset - the first value offset to write.

length - the number of values to write

Throws:

IOException

createRowIndexEntry

void createRowIndexEntry()
                  throws IOException

Create a row index entry at the current point in the stripe.

Throws:

IOException

flushStreams

void flushStreams()
           throws IOException

Flush the TreeWriter stream

Throws:

IOException

writeStripe

void writeStripe(OrcProto.StripeFooter.Builder stripeFooter,
                 OrcProto.StripeStatistics.Builder stats,
                 int requiredIndexEntries)
          throws IOException

Write the stripe out to the file.

Parameters:

stripeFooter - the stripe footer that contains the information about the layout of the stripe. The TreeWriterBase is required to update the footer with its information.

stats - the stripe statistics information

requiredIndexEntries - the number of index entries that are required. this is to check to make sure the row index is well formed.

Throws:

IOException

updateFileStatistics

void updateFileStatistics(OrcProto.StripeStatistics stripeStatistics)

During a stripe append, we need to update the file statistics.

Parameters:

stripeStatistics - the statistics for the new stripe

writeFileStatistics

void writeFileStatistics(OrcProto.Footer.Builder footer)

Add the file statistics to the file footer.

Parameters:

footer - the file footer builder