simpledb

Class TableStats

java.lang.Object

simpledb.TableStats

public class TableStats
extends java.lang.Object

TableStats represents statistics (e.g., histograms) about base tables in a query. This class is not needed in implementing proj1 and proj2.

Field Summary

Fields

Modifier and Type	Field and Description
private HeapFile	file
private int	ioCostPerPage
(package private) static int	IOCOSTPERPAGE
private DbFileIterator	iterator
private Tuple	maxTuple
private Tuple	minTuple
(package private) static int	NUM_HIST_BINS Number of bins for the histogram.
private int	numTuples
private static java.util.concurrent.ConcurrentHashMap<java.lang.String,TableStats>	statsMap
private int	tableId

Constructor Summary

Constructors

Constructor and Description
TableStats(int tableid, int ioCostPerPage) Create a new TableStats object, that keeps track of statistics on each column of a table

Method Summary

Modifier and Type	Method and Description
double	avgSelectivity(int field, Predicate.Op op) The average selectivity of the field under op.
static void	computeStatistics()
double	estimateScanCost() Estimates the cost of sequentially scanning the file, given that the cost to read a page is costPerPageIO.
double	estimateSelectivity(int field, Predicate.Op op, Field constant) Estimate the selectivity of predicate field op constant on the table.
int	estimateTableCardinality(double selectivityFactor) This method returns the number of tuples in the relation, given that a predicate with selectivity selectivityFactor is applied.
Tuple	getMaxTuple()
Tuple	getMinTuple()
static java.util.Map<java.lang.String,TableStats>	getStatsMap()
static TableStats	getTableStats(java.lang.String tablename)
static void	setStatsMap(java.util.HashMap<java.lang.String,TableStats> s)
static void	setTableStats(java.lang.String tablename, TableStats stats)
int	totalTuples()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

statsMap

private static final java.util.concurrent.ConcurrentHashMap<java.lang.String,TableStats> statsMap

IOCOSTPERPAGE

static final int IOCOSTPERPAGE

See Also:

Constant Field Values

NUM_HIST_BINS

static final int NUM_HIST_BINS

Number of bins for the histogram. Feel free to increase this value over 100, though our tests assume that you have at least 100 bins in your histograms.

See Also:

Constant Field Values

tableId

private int tableId

ioCostPerPage

private int ioCostPerPage

file

private HeapFile file

iterator

private DbFileIterator iterator

minTuple

private Tuple minTuple

maxTuple

private Tuple maxTuple

numTuples

private int numTuples

Constructor Detail

TableStats

public TableStats(int tableid,
                  int ioCostPerPage)

Create a new TableStats object, that keeps track of statistics on each column of a table

Parameters:

tableid - The table over which to compute statistics

ioCostPerPage - The cost per page of IO. This doesn't differentiate between sequential-scan IO and disk seeks.

Method Detail

getTableStats

public static TableStats getTableStats(java.lang.String tablename)

setTableStats

public static void setTableStats(java.lang.String tablename,
                                 TableStats stats)

setStatsMap

public static void setStatsMap(java.util.HashMap<java.lang.String,TableStats> s)

getStatsMap

public static java.util.Map<java.lang.String,TableStats> getStatsMap()

computeStatistics

public static void computeStatistics()

estimateScanCost

public double estimateScanCost()

Estimates the cost of sequentially scanning the file, given that the cost to read a page is costPerPageIO. You can assume that there are no seeks and that no pages are in the buffer pool. Also, assume that your hard drive can only read entire pages at once, so if the last page of the table only has one tuple on it, it's just as expensive to read as a full page. (Most real hard drives can't efficiently address regions smaller than a page at a time.)

Returns:

The estimated cost of scanning the table.

estimateTableCardinality

public int estimateTableCardinality(double selectivityFactor)

This method returns the number of tuples in the relation, given that a predicate with selectivity selectivityFactor is applied.

Parameters:

selectivityFactor - The selectivity of any predicates over the table

Returns:

The estimated cardinality of the scan with the specified selectivityFactor

avgSelectivity

public double avgSelectivity(int field,
                             Predicate.Op op)

The average selectivity of the field under op.

Parameters:

field - the index of the field

op - the operator in the predicate The semantic of the method is that, given the table, and then given a tuple, of which we do not know the value of the field, return the expected selectivity. You may estimate this value from the histograms.

Returns:

The averege selectivity

estimateSelectivity

public double estimateSelectivity(int field,
                                  Predicate.Op op,
                                  Field constant)

Estimate the selectivity of predicate field op constant on the table.

Parameters:

field - The field over which the predicate ranges

op - The logical operation in the predicate

constant - The value against which the field is compared

Returns:

The estimated selectivity (fraction of tuples that satisfy) the predicate

totalTuples

public int totalTuples()

Returns:

return the total number of tuples in this table

getMinTuple

public Tuple getMinTuple()

getMaxTuple

public Tuple getMaxTuple()