com.vertica.sdk.PartitionWriter Class Reference

Inheritance diagram for com.vertica.sdk.PartitionWriter:

Collaboration diagram for com.vertica.sdk.PartitionWriter:

Public Member Functions
	PartitionWriter (int nargs)
void	addCol (ByteBuffer arg, int colstride, VerticaType dt, String colName)
void	addCol (ByteBuffer arg, int colstride, VerticaType dt)
void	addCol (ByteBuffer arg, ByteBuffer da, int colstride, VerticaType dt)
void	addCol (ByteBuffer arg, ByteBuffer da, int colstride, VerticaType dt, String colName)
boolean	checkStringUserBlockInfo (int idx)
boolean	checkTimeUserBlockInfo (int idx)
void	copyFromInput (int dstIdx, PartitionReader input_reader, int srcIdx) throws UdfException
ByteBuffer	getColDataAreaRef (int idx)
ByteBuffer	getColRef (int idx)
TypeCoercion	getDefaultTypeCoercion ()
int	getNumCols ()
VNumeric	getNumericWriter (int idx)
int	getNumRows ()
BaseDataOID	getProcessBlockUserInfoVectorAt (int i)
VString	getStringWriter (int idx)
SizedColumnTypes	getTypeMetaData ()
abstract boolean	getWriteableBlock () throws UdfException, DestroyInvocation
boolean	next () throws UdfException, DestroyInvocation
void	resetBuffers ()
void	setBoolean (int idx, boolean r)
void	setBooleanNull (int idx)
void	setBooleanValue (int idx, Boolean value)
void	setDate (int idx, java.sql.Date r)
void	setDateNull (int idx)
void	setDateValue (int idx, java.sql.Date value)
TypeCoercion	setDefaultTypeCoercion (TypeCoercion coerce)
void	setDouble (int idx, double r)
void	setDoubleNull (int idx)
void	setDoubleValue (int idx, Double value)
void	setLong (int idx, long r)
void	setLongNull (int idx)
void	setLongValue (int idx, Long value)
void	setNumeric (int idx, BigDecimal bd)
void	setNumericNull (int idx)
void	setNumericValue (int idx, BigDecimal value)
void	setRowFromMap (Map rowData)
void	setRowFromMap (Map rowData, TypeCoercion coerce)
void	setString (int idx, String r)
void	setStringBytes (int idx, ByteBuffer bytes)
void	setStringNull (int idx)
void	setStringValue (int idx, String value)
void	setTimestamp (int idx, java.sql.Timestamp r)
void	setTimestampInfiniteNeg (int idx)
void	setTimestampInfinitePos (int idx)
void	setTimestampNull (int idx)
void	setTimestampValue (int idx, java.sql.Timestamp value)
void	setValue (int idx, Object value, TypeCoercion coerce)
void	setVMap (int idx, Map data)
void	setVMap (int idx, Map data, TypeCoercion coerce)
void	throwInCorrectUsageError (int idx)
	: Throwing UdfException for incorrect setter usage

Public Attributes
int	count
int	index
int	ncols
int	nrows
SizedColumnTypes	typeMetaData

Protected Member Functions
void	clear ()
ByteBuffer	getInlineColBuffer (int idx)
long	getRowSize (SizedColumnTypes types)
void	setRowCount (SizedColumnTypes types)

Static Protected Member Functions
static long	blockSizeGivenRowSize (long row_size)

Protected Attributes
ArrayList	coldataareas
ArrayList	cols
ArrayList	colstrides
ArrayList	currentPos

Detailed Description

PartitionWriter provides an iterator-based write interface over output data for a single partition. Automatically makes space a block-at-a-time, as needed.

Member Function Documentation

void com.vertica.sdk.VerticaBlock.addCol(ByteBuffer arg, int colstride, VerticaType dt, String colName)

inherited Add the location for reading a particular argument.

Parameters

arg: The base location to find data.
colstride: The stride between data instances.
dt: The type of input.
colName: Name of the column

Referenced by com.vertica.sdk.VerticaBlock.addCol().

void com.vertica.sdk.VerticaBlock.addCol(ByteBuffer arg, int colstride, VerticaType dt)

inherited Add the location for reading a particular argument.

Parameters

arg: The base location to find data.
colstride: The stride between data instances.
dt: The type of input.

void com.vertica.sdk.VerticaBlock.addCol(ByteBuffer arg, ByteBuffer da, int colstride, VerticaType dt)

inherited Add the location for reading a particular argument.

Parameters

arg: The base location to find data.
da: The location to find out of band string data.
colstride: The stride between data instances.
dt: The type of input.

void com.vertica.sdk.VerticaBlock.addCol(ByteBuffer arg, ByteBuffer da, int colstride, VerticaType dt, String colName)

inherited Add the location for reading a particular argument.

Parameters

arg: The base location to find data.
da: The location to find out of band string data.
colstride: The stride between data instances.
dt: The type of input.
colName: Name of the column

void com.vertica.sdk.PartitionWriter.copyFromInput(int dstIdx, PartitionReader input_reader, int srcIdx) throws UdfException

Copies a column from the input reader to the output writer. The data types and sizes of the source and destination columns must match exactly.

Parameters

dstIdx: The destination column index (in the output writer)
input_reader: The input reader from which to copy a column
srcIdx: The source column index (in the input reader)

ByteBuffer com.vertica.sdk.VerticaBlock.getColDataAreaRef(int idx)

inherited Get the ByteBuffer that stores out of line string data (Data Area) for the idx'th argument

Parameters

idx: The column number to retrieve from the input row.

Returns the ByteBuffer

Referenced by com.vertica.sdk.BlockReader.getVString().

ByteBuffer com.vertica.sdk.VerticaBlock.getColRef(int idx)

inherited Example

ByteBuffer a = arg_reader.getColPtr(0);

Parameters

idx: The column number to retrieve from the input row.

Returns a ByteBuffer to the idx'th argument, containing data for the column

Referenced by

TypeCoercion com.vertica.sdk.PartitionWriter.getDefaultTypeCoercion()

Give the default TypeCoercion instance, creating one if necessary. This may be useful in cases where the calling code would like to change the TypeCoercionPolicy or do TypeCoercion in the same manner as this PartitionWriter.

If a TypeCoercion is created, it is an instance of StandardTypeCoercion.

See Also #setRowFromMap(Map)

Parameters

coerce: A TypeCoercion instance.

Returns the former default, which may be null.

void com.vertica.sdk.PartitionWriter.setDoubleValue(int idx, Double value)

Set a Double value for column at index idx. Accepts null values.

Parameters

idx: column number
value: value to set it to

void com.vertica.sdk.PartitionWriter.setLongValue(int idx, Long value)

Set a Long value for column at index idx. Accepts null values.

Parameters

idx: column number
value: value to set it to

void com.vertica.sdk.PartitionWriter.setNumericValue(int idx, BigDecimal value)

Set a Numeric value for column at index idx. Accepts null values.

Parameters

idx: column number
value: value to set it to

void com.vertica.sdk.PartitionWriter.setRowFromMap(Map rowData)

Sets all columns on the current row according to the data provided, using the column names to match the data keys. See setRowFromMap(Map, TypeCoercion) for more details. This is a shortcut for using the default type coercion.

Parameters

rowData: The data. Null data is treated as empty.

See Also #setRowFromMap(Map, TypeCoercion)

getDefaultTypeCoercion()

void com.vertica.sdk.PartitionWriter.setRowFromMap(Map rowData, TypeCoercion coerce)

Sets all columns on the current row according to the data provided, using the column names to match the data keys.

For each column, TypeCoercion will be used to convert a value in rowData to the appropriate type for that column. This is done via setValue and may cause a TypeCoercion exception.

To find the corresponding value, the 'first' matching key (according to case-insensitive SQL identifier matching) in the map will be used. 'first' is not well defined, but is consistent.

If no corresponding value is found, the column will be set to null. This never raises an exception, regardless of the target column type. That is, FAIL_SDK_UNSUPPORTED_TYPE is ignored in that case.

In addition, if the target row structure appears to be a flex table, the necessary internal structures will be populated with an encoded vmap via setVMap(int, Map, TypeCoercion).

Parameters

rowData: The data. Null data is treated as empty.
coerce: The implementation of coercion

See Also #setVMap(int, Map, TypeCoercion)

setValue(int, Object, TypeCoercion)

void com.vertica.sdk.PartitionWriter.setStringBytes(int idx, ByteBuffer bytes)

Set a String value for column at index idx. Accepts null values.

Parameters

idx: The column index
bytes: A ByteBuffer with bytes between position and limit that are a string encoded according to Vertica's requirements. This buffer must fit within the column's length requirements.

Exceptions

IllegalArgumentException: if bytes is too long

void com.vertica.sdk.PartitionWriter.setStringValue(int idx, String value)

Set a String value for column at index idx. Accepts null values.

Parameters

idx: column number
value: value to set it to

void com.vertica.sdk.PartitionWriter.setTimestampValue(int idx, java.sql.Timestamp value)

Set a Timestamp value for column at index idx. Accepts null values.

Parameters

idx: column number
value: value to set it to

void com.vertica.sdk.PartitionWriter.setValue(int idx, Object value, TypeCoercion coerce)

Set a value for a column using type coercion. Type coercion attempts to generate an appropriate value for the target column type based on the input type.

The following target column types are supported by the SDK for this method. All others will generate a TypeCoercionException due to FAIL_SDK_UNSUPPORTED_TYPE if appropriate.

Bools
Integers (excluding Numeric)
Floats
String-like types (char and binary, along with var- and long variants)
Timestamp and timestamptz
Dates

Parameters

idx: The column index
value: The value to store in the column. May be null.
coerce: The type coercion policy.

Referenced by com.vertica.sdk.PartitionWriter.setRowFromMap().

void com.vertica.sdk.PartitionWriter.setVMap(int idx, Map data)

Encodes a VMap into the given column, which must be of binary type.

Parameters

idx: The index of the binary or varbinary column.
data: The data. Null data is treated as empty.

See Also #setVMap(int, Map, TypeCoercion)

getDefaultTypeCoercion()

void com.vertica.sdk.PartitionWriter.setVMap(int idx, Map data, TypeCoercion coerce)

Encodes a VMap into the given column, which must be of binary type. Uses standard type coercion to identify aggregate values and to serialize all values.

Parameters

idx: The index of the binary or varbinary column.
data: The data. Null data is treated as empty.
coerce: Ignored but still here for backwards compatibility.