org.apache.drill.exec.planner.index

Interface Statistics

All Known Implementing Classes:

MapRDBStatistics

public interface Statistics

Field Summary

Fields

Modifier and Type	Field and Description
static long	AVG_COLUMN_SIZE
static double	AVG_ROWSIZE_UNKNOWN
static double	ROWCOUNT_HUGE
static double	ROWCOUNT_UNKNOWN

Method Summary

Modifier and Type	Method and Description
String	buildUniqueIndexIdentifier(IndexDescriptor idx) Returns a unique index identifier
double	getAvgRowSize(String tabIdxName, boolean isIndexScan) Returns the average row size for the specified filter condition
double	getLeadingRowCount(org.apache.calcite.rex.RexNode condition, String tabIdxName, DrillScanRelBase scanRel) Returns the leading rowcount for the specified filter condition Leading rowcount means rowcount for filter condition only on leading index columns.
double	getRowCount(org.apache.calcite.rex.RexNode condition, String tabIdxName, org.apache.calcite.rel.RelNode scanRel) Returns the rowcount for the specified filter condition
boolean	initialize(org.apache.calcite.rex.RexNode condition, DrillScanRelBase scanRel, IndexCallContext context)
boolean	isStatsAvailable() Returns whether statistics are available.

Field Detail

ROWCOUNT_UNKNOWN

static final double ROWCOUNT_UNKNOWN

See Also:

Constant Field Values

ROWCOUNT_HUGE

static final double ROWCOUNT_HUGE

See Also:

Constant Field Values

AVG_ROWSIZE_UNKNOWN

static final double AVG_ROWSIZE_UNKNOWN

See Also:

Constant Field Values

AVG_COLUMN_SIZE

static final long AVG_COLUMN_SIZE

See Also:

Constant Field Values

Method Detail

isStatsAvailable

boolean isStatsAvailable()

Returns whether statistics are available. Should be called prior to using the statistics

buildUniqueIndexIdentifier

String buildUniqueIndexIdentifier(IndexDescriptor idx)

Returns a unique index identifier

Parameters:

idx - - Index specified as a IndexDescriptor

Returns:

The unique index identifier

getRowCount

double getRowCount(org.apache.calcite.rex.RexNode condition,
                   String tabIdxName,
                   org.apache.calcite.rel.RelNode scanRel)

Returns the rowcount for the specified filter condition

Parameters:

condition - - Filter specified as a RexNode

tabIdxName - - The index name generated using buildUniqueIndexIdentifier

scanRel - - The current scan rel

Returns:

the rowcount for the filter

getLeadingRowCount

double getLeadingRowCount(org.apache.calcite.rex.RexNode condition,
                          String tabIdxName,
                          DrillScanRelBase scanRel)

Returns the leading rowcount for the specified filter condition Leading rowcount means rowcount for filter condition only on leading index columns.

Parameters:

condition - - Filter specified as a RexNode

tabIdxName - - The index name generated using buildUniqueIndexIdentifier

scanRel - - The current scan rel

Returns:

the leading rowcount

getAvgRowSize

double getAvgRowSize(String tabIdxName,
                     boolean isIndexScan)

Returns the average row size for the specified filter condition

Parameters:

tabIdxName - - The index name generated using buildUniqueIndexIdentifier

isIndexScan - - Whether the current rel is an index scan (false for primary table)

initialize

boolean initialize(org.apache.calcite.rex.RexNode condition,
                   DrillScanRelBase scanRel,
                   IndexCallContext context)