ibis::bin Class Reference

The equality encoded bitmap index with binning. More...

#include <ibin.h>

Inheritance diagram for ibis::bin:

List of all members.

Public Types

typedef std::map< double,
granule * > granuleMap

Public Member Functions

long append (const array_t< uint32_t > &ind)

Append a list of integers representing bin numbers.

long append (const ibis::bin &tail)

Append the tail to this index.

virtual long append (const char *dt, const char *df, uint32_t nnew)

Extend the index.

bin (const ibis::column *c, const char *f, const std::vector< double > &bd)

bin (const ibis::column *c, const char *f, const array_t< double > &bd)

bin (const ibis::column *c, ibis::fileManager::storage *st, uint32_t offset=8)

bin (const ibis::column *c=0, const char *f=0)

Construct a bitmap index from current data.

bin (const ibis::bin &rhs)

virtual void binBoundaries (std::vector< double > &) const

The function binBoundaries and binWeights return bin boundaries and counts of each bin respectively.

virtual void binWeights (std::vector< uint32_t > &) const

long checkBin (const ibis::qRange &cmp, uint32_t jbin, const ibis::bitvector &mask, ibis::bitvector &res) const

Candidate check using the binned values.

long checkBin (const ibis::qRange &cmp, uint32_t jbin, ibis::bitvector &res) const

Candidate check using the binned values.

virtual int contractRange (ibis::qContinuousRange &rng) const

virtual int64_t estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, const ibis::bitvector &mask, const ibis::qRange *const range1, const ibis::qRange *const range2) const

virtual int64_t estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, const ibis::bitvector &mask) const

virtual int64_t estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr) const

virtual void estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, const ibis::bitvector &mask, const ibis::qRange *const range1, const ibis::qRange *const range2, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const

virtual void estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, const ibis::bitvector &mask, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const

virtual void estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const

Estimate the number of hits for nonsymmetric joins.

virtual int64_t estimate (const ibis::rangeJoin &expr, const ibis::bitvector &mask, const ibis::qRange *const range1, const ibis::qRange *const range2) const

virtual void estimate (const ibis::rangeJoin &expr, const ibis::bitvector &mask, const ibis::qRange *const range1, const ibis::qRange *const range2, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const

Evaluating a join condition with one (likely composite) index.

virtual void estimate (const ibis::rangeJoin &expr, const ibis::bitvector &mask, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const

virtual void estimate (const ibis::rangeJoin &expr, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const

Estimate the hits for symmetric joins.

virtual uint32_t estimate (const ibis::qContinuousRange &expr) const

Returns an upper bound on the number of hits.

virtual void estimate (const ibis::qContinuousRange &expr, ibis::bitvector &lower, ibis::bitvector &upper) const

Computes an approximation of hits as a pair of lower and upper bounds.

virtual double estimateCost (const ibis::qDiscreteRange &expr) const

virtual double estimateCost (const ibis::qContinuousRange &expr) const

Estimate the code of evaluate a range condition.

virtual long evaluate (const ibis::qContinuousRange &expr, ibis::bitvector &hits) const

To evaluate the exact hits.

virtual int expandRange (ibis::qContinuousRange &rng) const

The functions expandRange and contractRange expands or contracts the boundaries of a range condition so that the new range will have exact answers using the function estimate.

virtual long getCumulativeDistribution (std::vector< double > &bds, std::vector< uint32_t > &cts) const

Cumulative distribution of the data.

virtual long getDistribution (std::vector< double > &bbs, std::vector< uint32_t > &cts) const

Binned distribution of the data.

virtual double getMax () const

The maximum value recorded in the index.

virtual double getMin () const

The minimum value recorded in the index.

virtual double getSum () const

Compute the approximate sum of all the values indexed.

array_t< uint32_t > * indices (const ibis::bitvector &mask) const

virtual const char * name () const

Returns the name of the index, similar to the function type, but returns a string instead.

virtual uint32_t numBins () const

virtual void print (std::ostream &out) const

Prints human readable information.

int read (int fdes, uint32_t offset, const char *fname)

Read an ibis::bin embedded inside a file.

virtual int read (ibis::fileManager::storage *st)

Reconstructs an index from an array of bytes.

virtual int read (const char *idxfile)

Reconstructs an index from the named file.

virtual void speedTest (std::ostream &out) const

Time some logical operations and print out their speed.

virtual INDEX_TYPE type () const

Returns an index type identifier.

virtual float undecidable (const ibis::qContinuousRange &expr, ibis::bitvector &iffy) const

Mark the position of the rows that can not be decided with this index.

virtual int write (const char *dt) const

Save index to a file.

Protected Member Functions

void addBounds (double lbd, double rbd, uint32_t nbins, uint32_t eqw)

virtual void adjustLength (uint32_t nrows)

bin (const ibis::column *c, const uint32_t nbits, ibis::fileManager::storage *st, uint32_t offset=8)

A constructor to handle the common portion of multicomponent encodings.

template<typename E>

void binning (const array_t< E > &varr, const array_t< double > &bd)

template<typename E>

void binning (const array_t< E > &varr)

void binning (const char *f)

Read the data file and partition the values into bins according to the specified bin boundary.

void binning (const char *f, const array_t< double > &bd)

void binning (const char *f, const std::vector< double > &bd)

Generate bins according to the specified boundaries.

template<typename E>

void binningT (const char *fname)

Read the data file, partition the values, and write out the bin ordered data with .bin suffix.

long binOrder (const char *fname) const

template<typename E>

long binOrderT (const char *fname) const

Write bin-ordered values.

template<typename E>

long checkBin0 (const ibis::qRange &cmp, uint32_t jbin, ibis::bitvector &res) const

template<typename E>

long checkBin1 (const ibis::qRange &cmp, uint32_t jbin, const ibis::bitvector &mask, ibis::bitvector &res) const

virtual void clear ()

Clear the existing content.

virtual double computeSum () const

template<typename E>

void construct (const array_t< E > &varr)

void convertGranules (granuleMap &gmap)

void divideBitmaps (const std::vector< ibis::bitvector * > &bms, std::vector< unsigned > &parts) const

virtual void locate (const ibis::qContinuousRange &expr, uint32_t &cand0, uint32_t &cand1, uint32_t &hit0, uint32_t &hit1) const

virtual void locate (const ibis::qContinuousRange &expr, uint32_t &cand0, uint32_t &cand1) const

virtual uint32_t locate (const double &val) const

template<typename E>

void mapGranules (const array_t< E > &, granuleMap &gmap) const

uint32_t parseNbins () const

unsigned parsePrec () const

unsigned parseScale () const

void printGranules (std::ostream &out, const granuleMap &gmap) const

void readBinBoundaries (const char *name, uint32_t nb)

void scanAndPartition (const char *, unsigned, uint32_t nbins=0)

The optional argument nbins can either be set outside or set to be the return value of function parseNbins.

template<typename E>

void scanAndPartition (const array_t< E > &, unsigned)

template<typename E>

void setBoundaries (const array_t< E > &varr)

void setBoundaries (array_t< double > &bnds, const ibis::bin &idx1, const array_t< uint32_t > cnt1, const array_t< uint32_t > cnt0) const

void setBoundaries (array_t< double > &bnds, const ibis::bin &bin0) const

void setBoundaries (const char *f)

Set bin boundaries.

void swap (bin &rhs)

int write (int fptr) const

Protected Attributes

array_t< double > bounds

The nominal boundaries.

array_t< double > maxval

The maximal values in each bin.

array_t< double > minval

The minimal values in each bin.

uint32_t nobs

Number of bitvectors.

Friends

class ibis::ambit

class ibis::band

class ibis::mesa

class ibis::mesh

class ibis::pack

class ibis::pale

class ibis::range

class ibis::zone

Classes

struct granule

A data structure to assist the mapping of values to lower precisions. More...

Detailed Description

The equality encoded bitmap index with binning.

The exact bin boundary assignment is controlled by indexing options '<binning ... />'.

The 0th bit vector represents x < bounds[0]; The (nobs-1)st bit vector represents x >= bounds[nobs-2]; The ith bit vector represents bounds[i-1] <= x < bounds[i], (0 < i < nbos-1).

Member Function Documentation

void ibis::bin::binBoundaries ( std::vector< double > & ) const [virtual]

The function binBoundaries and binWeights return bin boundaries and counts of each bin respectively.

Reimplemented from ibis::index.

Reimplemented in ibis::range, ibis::mesa, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::egale, ibis::bak, and ibis::bak2.

Referenced by getCumulativeDistribution(), and getDistribution().

void ibis::bin::binning ( const char * f ) [protected]

Read the data file and partition the values into bins according to the specified bin boundary.

References ibis::bitvector::adjustSize(), ibis::index::bits, ibis::bitvector::bitsPerLiteral(), bounds, ibis::util::logger::buffer(), ibis::BYTE, ibis::CATEGORY, clear(), ibis::index::clear(), ibis::index::col, ibis::bitvector::copy(), ibis::index::dataFileName(), ibis::DOUBLE, ibis::bitvector::firstIndexSet(), ibis::FLOAT, ibis::fileManager::getFile(), ibis::gVerbose, ibis::bitvector::indexSet::indices(), ibis::fileManager::instance(), ibis::INT, ibis::bitvector::indexSet::isRange(), ibis::column::logMessage(), ibis::column::logWarning(), ibis::LONG, maxval, minval, ibis::bitvector::indexSet::nIndices(), ibis::part::nRows(), ibis::index::nrows, ibis::column::partition(), ibis::horometer::realTime(), array_t< T >::resize(), ibis::bitvector::set(), ibis::SHORT, ibis::bitvector::size(), array_t< T >::size(), ibis::horometer::start(), ibis::horometer::stop(), ibis::column::type(), ibis::UBYTE, ibis::UINT, ibis::ULONG, and ibis::USHORT.

void ibis::bin::binning	(	const char *	f,
		const std::vector< double > &	bd
	)			`[protected]`

Generate bins according to the specified boundaries.

This version of the binning function takes an external specified bin boundaries -- if the array is too small to be valid, it uses the default option.

Note:: This function does not attempt to clear the content of the current data structure, the caller is responsible for this task!

References array_t< T >::back(), bounds, ibis::BYTE, ibis::index::col, ibis::DOUBLE, ibis::FLOAT, ibis::INT, ibis::LONG, ibis::column::name(), ibis::part::nRows(), ibis::index::nrows, ibis::column::partition(), array_t< T >::push_back(), array_t< T >::resize(), setBoundaries(), ibis::SHORT, array_t< T >::size(), ibis::column::type(), ibis::UBYTE, ibis::UINT, ibis::ULONG, and ibis::USHORT.

Referenced by ibis::moins::append(), ibis::entre::append(), ibis::egale::append(), append(), and bin().

template<typename E>

void ibis::bin::binningT ( const char * fname ) [inline, protected]

Read the data file, partition the values, and write out the bin ordered data with .bin suffix.

References ibis::bitvector::adjustSize(), array_t< T >::back(), array_t< T >::begin(), ibis::index::bits, ibis::bitvector::bitsPerLiteral(), bounds, ibis::util::logger::buffer(), ibis::index::clear(), ibis::index::col, ibis::bitvector::copy(), ibis::index::dataFileName(), ibis::bitvector::firstIndexSet(), ibis::index::fname, ibis::fileManager::getFile(), ibis::gVerbose, ibis::bitvector::indexSet::indices(), ibis::fileManager::instance(), ibis::bitvector::indexSet::isRange(), ibis::column::logMessage(), ibis::column::logWarning(), maxval, minval, name(), ibis::bitvector::indexSet::nIndices(), ibis::part::nRows(), ibis::index::nrows, ibis::column::partition(), array_t< T >::push_back(), ibis::horometer::realTime(), array_t< T >::resize(), ibis::bitvector::set(), ibis::bitvector::size(), array_t< T >::size(), ibis::horometer::start(), and ibis::horometer::stop().

long ibis::bin::checkBin	(	const ibis::qRange &	cmp,
		uint32_t	jbin,
		const ibis::bitvector &	mask,
		ibis::bitvector &	res
	)			const

Candidate check using the binned values.

The bitvector mask marks the actual values in the bin (because the bitmaps stored in bits do not directly corresponds to the bin).

References ibis::BYTE, ibis::bitvector::clear(), ibis::bitvector::cnt(), ibis::index::col, ibis::horometer::CPUTime(), ibis::DOUBLE, ibis::FLOAT, ibis::gVerbose, ibis::INT, ibis::column::logMessage(), ibis::column::logWarning(), ibis::LONG, ibis::column::name(), ibis::index::nrows, ibis::horometer::realTime(), ibis::SHORT, ibis::bitvector::size(), ibis::horometer::start(), ibis::horometer::stop(), ibis::column::type(), ibis::UBYTE, ibis::UINT, ibis::ULONG, and ibis::USHORT.

long ibis::bin::checkBin	(	const ibis::qRange &	cmp,
		uint32_t	jbin,
		ibis::bitvector &	res
	)			const

Candidate check using the binned values.

Returns the number of hits if successful, otherwise it returns a negative value.

References ibis::index::activate(), ibis::index::bits, ibis::BYTE, ibis::bitvector::clear(), ibis::index::col, ibis::horometer::CPUTime(), ibis::DOUBLE, ibis::FLOAT, ibis::gVerbose, ibis::INT, ibis::column::logMessage(), ibis::column::logWarning(), ibis::LONG, ibis::column::name(), ibis::horometer::realTime(), ibis::SHORT, ibis::horometer::start(), ibis::horometer::stop(), ibis::column::type(), ibis::UBYTE, ibis::UINT, ibis::ULONG, and ibis::USHORT.

Referenced by ibis::range::evaluate(), ibis::mesa::evaluate(), and evaluate().

void ibis::bin::clear ( ) [protected, virtual]

Clear the existing content.

Free the objectes pointed to by the pointers.

Reimplemented from ibis::index.

Reimplemented in ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::fuge, and ibis::egale.

References ibis::index::clear(), array_t< T >::clear(), maxval, and minval.

Referenced by ibis::bak2::append(), ibis::bak::append(), append(), bin(), binning(), ibis::zone::clear(), ibis::pale::clear(), ibis::pack::clear(), ibis::ambit::clear(), ibis::egale::clear(), ibis::fuge::clear(), ibis::mesa::mesa(), ibis::range::read(), ibis::bak2::read(), ibis::bak::read(), and read().

void ibis::bin::estimate	(	const ibis::rangeJoin &	expr,
		ibis::bitvector64 &	lower,
		ibis::bitvector64 &	upper
	)			const `[virtual]`

Estimate the hits for symmetric joins.

Evaluate the range join condition using the ibis::bin index.

Record the definite hits in lower, and all possible hits in upper. NOTE: upper includes all entries in lower.

References ibis::index::activate(), ibis::bitvector64::clear(), ibis::bitvector64::cnt(), ibis::index::col, ibis::horometer::CPUTime(), ibis::compRange::term::eval(), ibis::rangeJoin::getRange(), ibis::gVerbose, ibis::horometer::realTime(), ibis::bitvector64::set(), ibis::bitvector64::size(), ibis::horometer::start(), ibis::horometer::stop(), and ibis::compRange::term::termType().

void ibis::bin::estimate	(	const ibis::qContinuousRange &	expr,
		ibis::bitvector &	lower,
		ibis::bitvector &	upper
	)			const `[virtual]`

Computes an approximation of hits as a pair of lower and upper bounds.

Parameters:

	expr	the query expression to be evaluated.
	lower	a bitvector marking a subset of the hits. All rows marked with one (1) are definitely hits.
	upper	a bitvector marking a superset of the hits. All hits are marked with one, but some of the rows marked one may not be hits. If the variable upper is empty, the variable lower is assumed to contain the exact answer.

Reimplemented from ibis::index.

Reimplemented in ibis::range, ibis::mesa, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::fuge, ibis::egale, ibis::moins, and ibis::entre.

References ibis::index::activate(), ibis::index::bits, ibis::bitvector::clear(), ibis::bitvector::copy(), ibis::gVerbose, ibis::bitvector::set(), and ibis::index::sumBits().

Referenced by estimate().

long ibis::bin::evaluate	(	const ibis::qContinuousRange &	expr,
		ibis::bitvector &	hits
	)			const `[virtual]`

To evaluate the exact hits.

On success, return the number of hits, otherwise a negative value is returned.

Implements ibis::index.

Reimplemented in ibis::range, ibis::mesa, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::fuge, ibis::egale, ibis::moins, and ibis::entre.

References ibis::index::activate(), ibis::index::bits, checkBin(), ibis::bitvector::cnt(), ibis::index::col, ibis::bitvector::copy(), ibis::part::doScan(), ibis::gVerbose, ibis::column::logWarning(), ibis::column::partition(), ibis::bitvector::set(), ibis::bitvector::size(), ibis::index::sumBits(), and ibis::bitvector::swap().

int ibis::bin::expandRange ( ibis::qContinuousRange & rng ) const [virtual]

The functions expandRange and contractRange expands or contracts the boundaries of a range condition so that the new range will have exact answers using the function estimate.

The default implementation provided does nothing since this is only meaningful for indices based on bins.

Reimplemented from ibis::index.

Reimplemented in ibis::range, ibis::bak, and ibis::bak2.

References ibis::qContinuousRange::leftBound(), ibis::qContinuousRange::leftOperator(), minval, ibis::qContinuousRange::rightBound(), ibis::qContinuousRange::rightOperator(), and array_t< T >::size().

long ibis::bin::getCumulativeDistribution	(	std::vector< double > &	bds,
		std::vector< uint32_t > &	cts
	)			const `[virtual]`

Cumulative distribution of the data.

A brute-force approach to get an accurate cumulative distribution.

Reimplemented from ibis::index.

References binBoundaries(), ibis::util::logger::buffer(), ibis::index::col, ibis::column::logMessage(), and ibis::column::upperBound().

long ibis::bin::getDistribution	(	std::vector< double > &	bbs,
		std::vector< uint32_t > &	cts
	)			const `[virtual]`

Binned distribution of the data.

A brute-force approach to get an accurate distribution.

Reimplemented from ibis::index.

References binBoundaries(), ibis::util::logger::buffer(), ibis::index::col, and ibis::column::logMessage().

double ibis::bin::getSum ( ) const [virtual]

Compute the approximate sum of all the values indexed.

If it decides that computing the sum directly from the vertical partition is more efficient, it will return NaN immediately.

Reimplemented from ibis::index.

Reimplemented in ibis::range, ibis::mesa, ibis::ambit, ibis::pack, ibis::egale, ibis::moins, and ibis::entre.

References ibis::index::col, ibis::column::elementSize(), ibis::part::nRows(), and ibis::column::partition().

virtual const char* ibis::bin::name ( ) const [inline, virtual]

Returns the name of the index, similar to the function type, but returns a string instead.

Implements ibis::index.

Reimplemented in ibis::range, ibis::mesa, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::fuge, ibis::egale, ibis::moins, ibis::entre, ibis::bak, and ibis::bak2.

Referenced by binningT(), binOrderT(), and read().

void ibis::bin::print ( std::ostream & out ) const [virtual]

Prints human readable information.

Outputs information about the index as text to the specified output stream.

Implements ibis::index.

Reimplemented in ibis::range, ibis::mesa, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::fuge, ibis::egale, ibis::moins, ibis::entre, ibis::bak, and ibis::bak2.

References ibis::index::bits, ibis::index::col, ibis::gVerbose, maxval, minval, ibis::column::name(), ibis::part::name(), and ibis::column::partition().

Referenced by append(), and bin().

int ibis::bin::read	(	int	fdes,
		uint32_t	start,
		const char *	fn
	)

Read an ibis::bin embedded inside a file.

Read from a file starting from an arbitrary offset.

This is intended to be used by multi-level indices.

Reimplemented in ibis::range.

References ibis::index::bits, bounds, clear(), ibis::index::col, ibis::index::fname, ibis::fileManager::instance(), ibis::column::logWarning(), maxval, minval, ibis::index::nrows, ibis::index::offsets, ibis::fileManager::pageSize(), ibis::fileManager::recordPages(), ibis::bitvector::setSize(), ibis::bitvector::size(), ibis::index::str, and array_t< T >::swap().

int ibis::bin::read ( ibis::fileManager::storage * st ) [virtual]

Reconstructs an index from an array of bytes.

Intended for internal use only!

Implements ibis::index.

Reimplemented in ibis::range, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::fuge, and ibis::egale.

References ibis::fileManager::storage::begin(), ibis::index::bits, bounds, clear(), ibis::index::col, array_t< T >::copy(), ibis::index::fname, ibis::fileManager::storage::isFileMap(), ibis::column::logWarning(), maxval, minval, ibis::index::nrows, ibis::index::offsets, ibis::bitvector::setSize(), ibis::bitvector::size(), ibis::index::str, and array_t< T >::swap().

int ibis::bin::read ( const char * name ) [virtual]

Reconstructs an index from the named file.

The name can be the directory containing an index file. In this case, the name of the index file must be the name of the column followed by ".idx" suffix.

Implements ibis::index.

Reimplemented in ibis::range, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::fuge, ibis::egale, ibis::bak, and ibis::bak2.

References ibis::index::bits, bounds, clear(), ibis::index::col, ibis::index::fname, ibis::gVerbose, ibis::index::indexFileName(), ibis::fileManager::instance(), ibis::column::logMessage(), ibis::column::logWarning(), maxval, minval, name(), ibis::index::nrows, ibis::index::offsets, ibis::fileManager::pageSize(), ibis::fileManager::recordPages(), ibis::bitvector::setSize(), ibis::bitvector::size(), ibis::index::str, and array_t< T >::swap().

Referenced by bin(), ibis::zone::read(), ibis::pale::read(), ibis::pack::read(), ibis::fuge::read(), ibis::ambit::read(), ibis::range::read(), ibis::bak2::read(), and ibis::bak::read().

void ibis::bin::scanAndPartition	(	const char *	f,
		unsigned	eqw,
		uint32_t	nbins = `0`
	)			`[protected]`

The optional argument nbins can either be set outside or set to be the return value of function parseNbins.

References array_t< T >::back(), array_t< T >::begin(), ibis::util::logger::buffer(), array_t< T >::clear(), ibis::index::col, ibis::index::divideCounts(), ibis::DOUBLE, array_t< T >::end(), ibis::FLOAT, ibis::part::nRows(), ibis::column::partition(), array_t< T >::push_back(), array_t< T >::reserve(), array_t< T >::size(), ibis::column::type(), and ibis::column::upperBound().

void ibis::bin::setBoundaries ( const char * f ) [protected]

Set bin boundaries.

Parse the index specification to determine the bin boundaries and store the result in member variable bounds.

The bin specification can be of the following, where all fields are optional.

```
 equal([_-]?)[weight|length|ratio]) 
```

 no=xxx|nbins=xxx|bins:(\[begin, end, no=xxx\))+

 <binning (start=begin end=end nbins=xxx scale=[linear|log])* />

 <binning binFile=file-name[, nbins=xxx] />

The bin speficication can be read from the column object, the table object containing the column, or the global ibis::gParameters object under the name of table-name.column-name.index. If no index specification is found, it builts approximate equal weight bins.

Note:: If equal weight is specified, it take precedence over all other specification.

References ibis::util::logger::buffer(), ibis::index::col, ibis::DOUBLE, ibis::FLOAT, ibis::gVerbose, ibis::column::indexSpec(), ibis::column::logMessage(), ibis::column::logWarning(), ibis::column::lowerBound(), ibis::column::type(), and ibis::column::upperBound().

Referenced by ibis::moins::append(), ibis::entre::append(), ibis::egale::append(), append(), bin(), and binning().

float ibis::bin::undecidable	(	const ibis::qContinuousRange &	expr,
		ibis::bitvector &	iffy
	)			const `[virtual]`

Mark the position of the rows that can not be decided with this index.

Parameters:

	expr	the range conditions to be evaluated.
	iffy	the bitvector marking the positions of rows that can not be decided using the index. Return value is the expected fraction of undecided rows that might satisfy the range conditions.

Reimplemented from ibis::index.

Reimplemented in ibis::range, ibis::mesa, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, and ibis::egale.

References ibis::index::activate(), ibis::index::bits, ibis::bitvector::copy(), ibis::qContinuousRange::leftBound(), minval, ibis::qContinuousRange::rightBound(), and ibis::bitvector::set().

int ibis::bin::write ( const char * name ) const [virtual]

Save index to a file.

Outputs the index in a compact binary format to the named file or directory. The index file contains a header that can be identified by the function isIndex.

Implements ibis::index.

Reimplemented in ibis::range, ibis::mesa, ibis::ambit, ibis::pale, ibis::pack, ibis::zone, ibis::fuge, ibis::egale, ibis::moins, ibis::entre, ibis::bak, and ibis::bak2.

References ibis::index::activate(), array_t< T >::back(), array_t< T >::begin(), ibis::index::BINNING, ibis::index::bits, ibis::index::col, ibis::index::fname, ibis::gVerbose, ibis::index::indexFileName(), ibis::column::logMessage(), ibis::column::logWarning(), maxval, and minval.

Referenced by append(), and ibis::fuge::write().

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

ibin.h
ibin.cpp


Public Types
typedef std::map< double, granule * >	granuleMap
Public Member Functions
long	append (const array_t< uint32_t > &ind)
	Append a list of integers representing bin numbers.
long	append (const ibis::bin &tail)
	Append the `tail` to this index.
virtual long	append (const char dt, const char df, uint32_t nnew)
	Extend the index.
	bin (const ibis::column c, const char f, const std::vector< double > &bd)
	bin (const ibis::column c, const char f, const array_t< double > &bd)
	bin (const ibis::column c, ibis::fileManager::storage st, uint32_t offset=8)
	bin (const ibis::column c=0, const char f=0)
	Construct a bitmap index from current data.
	bin (const ibis::bin &rhs)
virtual void	binBoundaries (std::vector< double > &) const
	The function binBoundaries and binWeights return bin boundaries and counts of each bin respectively.
virtual void	binWeights (std::vector< uint32_t > &) const
long	checkBin (const ibis::qRange &cmp, uint32_t jbin, const ibis::bitvector &mask, ibis::bitvector &res) const
	Candidate check using the binned values.
long	checkBin (const ibis::qRange &cmp, uint32_t jbin, ibis::bitvector &res) const
	Candidate check using the binned values.
virtual int	contractRange (ibis::qContinuousRange &rng) const
virtual int64_t	estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, const ibis::bitvector &mask, const ibis::qRange const range1, const ibis::qRange const range2) const
virtual int64_t	estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, const ibis::bitvector &mask) const
virtual int64_t	estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr) const
virtual void	estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, const ibis::bitvector &mask, const ibis::qRange const range1, const ibis::qRange const range2, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const
virtual void	estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, const ibis::bitvector &mask, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const
virtual void	estimate (const ibis::bin &idx2, const ibis::rangeJoin &expr, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const
	Estimate the number of hits for nonsymmetric joins.
virtual int64_t	estimate (const ibis::rangeJoin &expr, const ibis::bitvector &mask, const ibis::qRange const range1, const ibis::qRange const range2) const
virtual void	estimate (const ibis::rangeJoin &expr, const ibis::bitvector &mask, const ibis::qRange const range1, const ibis::qRange const range2, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const
	Evaluating a join condition with one (likely composite) index.
virtual void	estimate (const ibis::rangeJoin &expr, const ibis::bitvector &mask, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const
virtual void	estimate (const ibis::rangeJoin &expr, ibis::bitvector64 &lower, ibis::bitvector64 &upper) const
	Estimate the hits for symmetric joins.
virtual uint32_t	estimate (const ibis::qContinuousRange &expr) const
	Returns an upper bound on the number of hits.
virtual void	estimate (const ibis::qContinuousRange &expr, ibis::bitvector &lower, ibis::bitvector &upper) const
	Computes an approximation of hits as a pair of lower and upper bounds.
virtual double	estimateCost (const ibis::qDiscreteRange &expr) const
virtual double	estimateCost (const ibis::qContinuousRange &expr) const
	Estimate the code of evaluate a range condition.
virtual long	evaluate (const ibis::qContinuousRange &expr, ibis::bitvector &hits) const
	To evaluate the exact hits.
virtual int	expandRange (ibis::qContinuousRange &rng) const
	The functions expandRange and contractRange expands or contracts the boundaries of a range condition so that the new range will have exact answers using the function estimate.
virtual long	getCumulativeDistribution (std::vector< double > &bds, std::vector< uint32_t > &cts) const
	Cumulative distribution of the data.
virtual long	getDistribution (std::vector< double > &bbs, std::vector< uint32_t > &cts) const
	Binned distribution of the data.
virtual double	getMax () const
	The maximum value recorded in the index.
virtual double	getMin () const
	The minimum value recorded in the index.
virtual double	getSum () const
	Compute the approximate sum of all the values indexed.
array_t< uint32_t > *	indices (const ibis::bitvector &mask) const
virtual const char *	name () const
	Returns the name of the index, similar to the function `type`, but returns a string instead.
virtual uint32_t	numBins () const
virtual void	print (std::ostream &out) const
	Prints human readable information.
int	read (int fdes, uint32_t offset, const char *fname)
	Read an ibis::bin embedded inside a file.
virtual int	read (ibis::fileManager::storage *st)
	Reconstructs an index from an array of bytes.
virtual int	read (const char *idxfile)
	Reconstructs an index from the named file.
virtual void	speedTest (std::ostream &out) const
	Time some logical operations and print out their speed.
virtual INDEX_TYPE	type () const
	Returns an index type identifier.
virtual float	undecidable (const ibis::qContinuousRange &expr, ibis::bitvector &iffy) const
	Mark the position of the rows that can not be decided with this index.
virtual int	write (const char *dt) const
	Save index to a file.
Protected Member Functions
void	addBounds (double lbd, double rbd, uint32_t nbins, uint32_t eqw)
virtual void	adjustLength (uint32_t nrows)
	bin (const ibis::column c, const uint32_t nbits, ibis::fileManager::storage st, uint32_t offset=8)
	A constructor to handle the common portion of multicomponent encodings.
template<typename E>
void	binning (const array_t< E > &varr, const array_t< double > &bd)
template<typename E>
void	binning (const array_t< E > &varr)
void	binning (const char *f)
	Read the data file and partition the values into bins according to the specified bin boundary.
void	binning (const char *f, const array_t< double > &bd)
void	binning (const char *f, const std::vector< double > &bd)
	Generate bins according to the specified boundaries.
template<typename E>
void	binningT (const char *fname)
	Read the data file, partition the values, and write out the bin ordered data with .bin suffix.
long	binOrder (const char *fname) const
template<typename E>
long	binOrderT (const char *fname) const
	Write bin-ordered values.
template<typename E>
long	checkBin0 (const ibis::qRange &cmp, uint32_t jbin, ibis::bitvector &res) const
template<typename E>
long	checkBin1 (const ibis::qRange &cmp, uint32_t jbin, const ibis::bitvector &mask, ibis::bitvector &res) const
virtual void	clear ()
	Clear the existing content.
virtual double	computeSum () const
template<typename E>
void	construct (const array_t< E > &varr)
void	convertGranules (granuleMap &gmap)
void	divideBitmaps (const std::vector< ibis::bitvector * > &bms, std::vector< unsigned > &parts) const
virtual void	locate (const ibis::qContinuousRange &expr, uint32_t &cand0, uint32_t &cand1, uint32_t &hit0, uint32_t &hit1) const
virtual void	locate (const ibis::qContinuousRange &expr, uint32_t &cand0, uint32_t &cand1) const
virtual uint32_t	locate (const double &val) const
template<typename E>
void	mapGranules (const array_t< E > &, granuleMap &gmap) const
uint32_t	parseNbins () const
unsigned	parsePrec () const
unsigned	parseScale () const
void	printGranules (std::ostream &out, const granuleMap &gmap) const
void	readBinBoundaries (const char *name, uint32_t nb)
void	scanAndPartition (const char *, unsigned, uint32_t nbins=0)
	The optional argument `nbins` can either be set outside or set to be the return value of function parseNbins.
template<typename E>
void	scanAndPartition (const array_t< E > &, unsigned)
template<typename E>
void	setBoundaries (const array_t< E > &varr)
void	setBoundaries (array_t< double > &bnds, const ibis::bin &idx1, const array_t< uint32_t > cnt1, const array_t< uint32_t > cnt0) const
void	setBoundaries (array_t< double > &bnds, const ibis::bin &bin0) const
void	setBoundaries (const char *f)
	Set bin boundaries.
void	swap (bin &rhs)
int	write (int fptr) const
Protected Attributes
array_t< double >	bounds
	The nominal boundaries.
array_t< double >	maxval
	The maximal values in each bin.
array_t< double >	minval
	The minimal values in each bin.
uint32_t	nobs
	Number of bitvectors.
Friends
class	ibis::ambit
class	ibis::band
class	ibis::mesa
class	ibis::mesh
class	ibis::pack
class	ibis::pale
class	ibis::range
class	ibis::zone
Classes
struct	granule
	A data structure to assist the mapping of values to lower precisions. More...