Public Member Functions
Uint32	getNoOfParentOperations () const
NdbQueryOperation *	getParentOperation (Uint32 parentNo) const
Uint32	getNoOfChildOperations () const
NdbQueryOperation *	getChildOperation (Uint32 childNo) const
const NdbQueryOperationDef &	getQueryOperationDef () const
NdbQuery &	getQuery () const
NdbRecAttr *	getValue (const char anAttrName, char resultBuffer=0)
NdbRecAttr *	getValue (Uint32 anAttrId, char *resultBuffer=0)
NdbRecAttr *	getValue (const NdbDictionary::Column column, char resultBuffer=0)
int	setResultRowBuf (const NdbRecord rec, char resBuffer, const unsigned char *result_mask=0)
int	setResultRowRef (const NdbRecord rec, const char &bufRef, const unsigned char *result_mask=0)
NdbQueryOperationImpl &	getImpl () const
int	setOrdering (NdbQueryOptions::ScanOrdering ordering)
NdbQueryOptions::ScanOrdering	getOrdering () const
int	setParallelism (Uint32 parallelism)
int	setMaxParallelism ()
int	setAdaptiveParallelism ()
int	setBatchSize (Uint32 batchSize)
int	setInterpretedCode (const NdbInterpretedCode &code) const
NdbQuery::NextResultOutcome	firstResult ()
NdbQuery::NextResultOutcome	nextResult (bool fetchAllowed=true, bool forceSend=false)
bool	isRowNULL () const
bool	isRowChanged () const

Friends
class	NdbQueryOperationImpl

Detailed Description

Definition at line 231 of file NdbQueryOperation.hpp.

Member Function Documentation

NdbQuery::NextResultOutcome NdbQueryOperation::firstResult ( )

Local cursor:

Navigate to first result row in this batch of results which depends on the current row(s) from all its ancestors.

Returns

NextResult_error (-1): if unsuccessful,
NextResult_gotRow (0): if another tuple was received, and
NextResult_scanComplete (1): if there are no more tuples to scan.
NextResult_bufferEmpty (2): if there are no more cached records in NdbApi

Definition at line 1530 of file NdbQueryOperation.cpp.

NdbQueryOperationImpl& NdbQueryOperation::getImpl ( ) const

inline

Get object implementing NdbQueryOperation interface.

Definition at line 322 of file NdbQueryOperation.hpp.

NdbQueryOptions::ScanOrdering NdbQueryOperation::getOrdering ( ) const

Get the result ordering for this operation.

Definition at line 1503 of file NdbQueryOperation.cpp.

NdbRecAttr * NdbQueryOperation::getValue	(	const char *	anAttrName,
		char *	resultBuffer = `0`
	)

Defines a retrieval operation of an attribute value. The NDB API allocate memory for the NdbRecAttr object that will hold the returned attribute value.

Note: Note that it is the applications responsibility to allocate enough memory for resultBuffer (if non-NULL). The buffer resultBuffer supplied by the application must be aligned appropriately. The buffer is used directly (avoiding a copy penalty) only if it is aligned on a 4-byte boundary and the attribute size in bytes (i.e. NdbRecAttr::attrSize times NdbRecAttr::arraySize is a multiple of 4).; There are three versions of NdbQueryOperation::getValue with slightly different parameters.; This method does not fetch the attribute value from the database! The NdbRecAttr object returned by this method is not readable/printable before the transaction has been executed with NdbTransaction::execute.

Parameters

anAttrName	Attribute name
resultBuffer	If this is non-NULL, then the attribute value will be returned in this parameter. If NULL, then the attribute value will only be stored in the returned NdbRecAttr object.

Returns: An NdbRecAttr object to hold the value of the attribute, or a NULL pointer (indicating error).

Definition at line 1450 of file NdbQueryOperation.cpp.

NdbQuery::NextResultOutcome NdbQueryOperation::nextResult	(	bool	fetchAllowed = `true`,
		bool	forceSend = `false`
	)

Local cursor:

Get the next tuple(s) from this operation (and all its descendants?) which depends on the current row(s) from all its ancestors.

Result row / columns will be updated in the respective result handlers as previously specified on each NdbQueryOperation either by assigning a NdbRecord/rowBuffer or assigning NdbRecAttr to each column to be retrieved.

If the set of cached records in the NdbApi has been consumed, more will be requested from the datanodes only iff:

This NdbOperation is the root of the entire pushed NdbQuery.
'fetchAllowed==true'

The arguments fetchAllowed and forceSend are ignored if this operation is not the root of the pushed query.

Returns

NextResult_error (-1): if unsuccessful,
NextResult_gotRow (0): if another tuple was received, and
NextResult_scanComplete (1): if there are no more tuples to scan.
NextResult_bufferEmpty (2): if there are no more cached records in NdbApi

Definition at line 1536 of file NdbQueryOperation.cpp.

Here is the call graph for this function:

int NdbQueryOperation::setAdaptiveParallelism ( )

Let the system dynamically choose the number of fragments to scan in parallel. The system will try to choose a value that gives optimal performance. This is the default for all scans but the root scan. This method only implemented for non-root scan operations.

Returns: 0 if ok, -1 in case of error (call getNdbError() for details.)

Definition at line 1516 of file NdbQueryOperation.cpp.

Here is the call graph for this function:

int NdbQueryOperation::setBatchSize ( Uint32 batchSize )

Set the batch size (max rows per batch) for this operation. This only applies to scan operations, as lookup operations always will have the same batch size as its parent operation, or 1 if it is the root operation.

Parameters

batchSize Batch size (in number of rows). A value of 0 means use the default batch size.

Returns: 0 if ok, -1 in case of error (call getNdbError() for details.)

Definition at line 1520 of file NdbQueryOperation.cpp.

Here is the call graph for this function:

int NdbQueryOperation::setInterpretedCode ( const NdbInterpretedCode & code ) const

Set the NdbInterpretedCode needed for defining a conditional filter (aka: predicate) for this operation. Might be used both on scan and lookup operations.

Typically, one would create NdbScanFilter and NdbInterpretedCode objects on the stack, e.g.: NdbInterpretedCode code(table); NdbScanFilter filter(code); filter.begin(); filter.ge(0, 5U); // Check if column 1 is greater of equal to 5. filter.end(); queryOp->setInterpretedCode(code);

Parameters

code	The interpreted code. This object is copied internally, meaning that 'code' may be destroyed as soon as this method returns.

Returns: 0 if ok, -1 in case of error (call getNdbError() for details.)

Definition at line 1524 of file NdbQueryOperation.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

int NdbQueryOperation::setMaxParallelism ( )

Set the number of fragments to be scanned in parallel to the maximum possible value. This is the default for the root scan operation.

Returns: 0 if ok, -1 in case of error (call getNdbError() for details.)

Definition at line 1512 of file NdbQueryOperation.cpp.

Here is the call graph for this function:

int NdbQueryOperation::setOrdering ( NdbQueryOptions::ScanOrdering ordering )

Define result ordering for ordered index scan. It is an error to call this method on an operation that is not a scan, or to call it if an ordering was already set on the operation defintion by calling NdbQueryOperationDef::setOrdering().

Parameters

ordering The desired ordering of results.

Returns: 0 if ok, -1 in case of error (call getNdbError() for details.)

Definition at line 1497 of file NdbQueryOperation.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

int NdbQueryOperation::setParallelism ( Uint32 parallelism )

Set the number of fragments to be scanned in parallel. This only applies to table scans and non-sorted scans of ordered indexes. This method is only implemented for then root scan operation.

Returns: 0 if ok, -1 in case of error (call getNdbError() for details.)

Definition at line 1508 of file NdbQueryOperation.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

int NdbQueryOperation::setResultRowBuf	(	const NdbRecord *	rec,
		char *	resBuffer,
		const unsigned char *	result_mask = `0`
	)

Retrieval of entire or partial rows may also be specified. For partial retrieval a bitmask should supplied.

The behaviour of mixing NdbRecord retrieval style with NdbRecAttr is is undefined - It should probably not be allowed.

Parameters

rec	Is a pointer to a NdbRecord specifying the byte layout of the result row.

Defines a buffer sufficient large to hold the result row.

Refers a pointer which will be updated to refer the current result row for this operand.

Parameters

result_mask defines as subset of attributes to read. The column is only affected if 'mask[attrId >> 3] & (1<<(attrId & 7))' is set

Returns: 0 on success, -1 otherwise (call getNdbError() for details).

Definition at line 1475 of file NdbQueryOperation.cpp.

Here is the call graph for this function:

The documentation for this class was generated from the following files:

storage/ndb/src/ndbapi/NdbQueryOperation.hpp
storage/ndb/src/ndbapi/NdbQueryOperation.cpp

Public Member Functions

Friends

Detailed Description

Member Function Documentation