mondrian.spi

Class SegmentColumn

java.lang.Object

mondrian.spi.SegmentColumn

All Implemented Interfaces:

Serializable

public class SegmentColumn
extends Object
implements Serializable

Constrained columns are part of the SegmentHeader and SegmentCache. They uniquely identify a constrained column within a segment. Each segment can have many constrained columns. Each column can be constrained by multiple values at once (similar to a SQL in() predicate).

They are immutable and serializable.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
String	columnExpression
long	valueCount
SortedSet<Comparable>	values

Constructor Summary

Constructors

Constructor and Description
SegmentColumn(String columnExpression, long valueCount, SortedSet<Comparable> valueList) Creates a SegmentColumn.

Method Summary

Modifier and Type	Method and Description
boolean	equals(Object obj)
String	getColumnExpression() Returns the column expression of this constrained column.
long	getValueCount() Returns the number of distinct values that occur for this column in the database.
SortedSet<Comparable>	getValues() Returns an array of predicate values for this column.
int	hashCode()
SegmentColumn	merge(SegmentColumn col) Merges this column with another resulting in another whose values are super set of both.

Methods inherited from class java.lang.Object

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

Field Detail

columnExpression

public final String columnExpression

valueCount

public final long valueCount

values

public final SortedSet<Comparable> values

Constructor Detail

SegmentColumn

public SegmentColumn(String columnExpression,
                     long valueCount,
                     SortedSet<Comparable> valueList)

Creates a SegmentColumn.

Parameters:

columnExpression - SQL expression for the column. Unique within the star schema, including accesses to the same physical column via different join paths.

valueCount - Number of distinct values of this column in the database. For these purposes, null is counted as a value. If there are N distinct values of the column, and we have a collection of segments that cover N values, then Mondrian assumes that it is safe to roll up.

valueList - List of values to constrain the column to, or null if unconstrained. Values must be Comparable and immutable. For example, Integer, Boolean, String or Double.

Method Detail

merge

public SegmentColumn merge(SegmentColumn col)

Merges this column with another resulting in another whose values are super set of both.

getColumnExpression

public String getColumnExpression()

Returns the column expression of this constrained column.

Returns:

A column expression.

getValues

public SortedSet<Comparable> getValues()

Returns an array of predicate values for this column.

Returns:

An array of object values.

equals

public boolean equals(Object obj)

Overrides:

equals in class Object

hashCode

public int hashCode()

Overrides:

hashCode in class Object

getValueCount

public long getValueCount()

Returns the number of distinct values that occur for this column in the database.

Mondrian uses this to know that it can safely combine multiple segments to roll up. For example, if for the "quarter" column, one segment has values {"Q1", "Q2"} and another has values {"Q3", "Q4"}, and Mondrian knows that there are 4 values, then it can roll up.

If this method returns a value that is too low, Mondrian may generate incorrect results. If you don't know the number of values, return -1.

Returns:

Number of distinct values, including null, that occur for this column