mondrian.olap.fun

Class FunUtil

java.lang.Object

org.eigenbase.xom.XMLUtil

org.eigenbase.xom.XOMUtil

mondrian.olap.Util

mondrian.olap.fun.FunUtil

Direct Known Subclasses:

CacheFunDef.CacheFunResolver, FunDefBase, MultiResolver

public class FunUtil
extends Util

FunUtil contains a set of methods useful within the mondrian.olap.fun package.

Since:

1.0

Author:

jhyde

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	FunUtil.DescendingValueComparator Compares cell values, so that larger values compare first.
static class	FunUtil.ObjIntPair<T> Tuple consisting of an object and an integer.
static class	FunUtil.OrderKey

Nested classes/interfaces inherited from class mondrian.olap.Util

Util.AbstractFlatList<T>, Util.ByteMatcher, Util.ErrorCellValue, Util.Flat2List<T>, Util.Flat3List<T>, Util.Functor1<RT,PT>, Util.GcIterator<T>, Util.MemoryInfo, Util.PropertyList, Util.SqlNullSafeComparator

Field Summary

Fields

Modifier and Type	Field and Description
static boolean	BooleanNull Null value in three-valued boolean logic.
static double	DoubleEmpty Special value which indicates that a double computation has returned the MDX EMPTY value.
static double	DoubleNull Special value which indicates that a double computation has returned the MDX null value.
static int	IntegerNull Special value which indicates that an int computation has returned the MDX null value.
static mondrian.olap.fun.FunUtil.NullMember	NullMember

Fields inherited from class mondrian.olap.Util

DEBUG, EmptyValue, IBM_JVM, JdbcVersion, JVM_INSTANCE_UUID, nl, nullValue, Retrowoven

Constructor Summary

Constructors

Constructor and Description
FunUtil()

Method Summary

Modifier and Type	Method and Description
static Object	avg(Evaluator evaluator, TupleList members, Calc calc)
static Double	box(double d) Converts a double (primitive) value to a Double.
static Integer	box(int n) Converts an int (primitive) value to an Integer.
static void	checkIterListResultStyles(Calc calc)
static void	checkListResultStyles(Calc calc)
static int	compareHierarchically(Member m1, Member m2, boolean post) Compares a pair of members according to their positions in a prefix-order (or postfix-order, if post is true) walk over a hierarchy.
static int	compareSiblingMembers(Member m1, Member m2) Compares two members which are known to have the same parent.
static int	compareValues(double d1, double d2) Compares double-precision values according to MDX semantics.
static int	compareValues(Object value0, Object value1) Compares two cell values.
static int	count(Evaluator evaluator, TupleIterable iterable, boolean includeEmpty)
static int	decodeCategory(String flags, int offset) Decodes the offsetth character of an encoded method signature into a type category.
static int[]	decodeParameterCategories(String flags) Decodes a string of parameter types into an array of type codes.
static int	decodeReturnCategory(String flags) Decodes the signature of a function into a category code which describes the return type of the operator.
static Syntax	decodeSyntacticType(String flags) Decodes the syntactic type of an operator.
static Hierarchy	getDimensionDefaultHierarchy(Dimension dimension) Returns the default hierarchy of a dimension, or null if there is no default.
static List<Member>	getNonEmptyMemberChildren(Evaluator evaluator, Member member)
static Map<Member,Access>	getNonEmptyMemberChildrenWithDetails(Evaluator evaluator, Member member)
static void	hierarchizeMemberList(List<Member> memberList, boolean post) Sorts a list of members into hierarchical order.
static TupleList	hierarchizeTupleList(TupleList tupleList, boolean post) Sorts a list of tuples into hierarchical order.
static Member[]	makeNullTuple(TupleType tupleType)
static Object	max(Evaluator evaluator, TupleList members, Calc exp)
static Object	min(Evaluator evaluator, TupleList members, Calc calc)
static RuntimeException	newEvalException(FunDef funDef, String message) Creates an exception which indicates that an error has occurred while executing a given function.
static RuntimeException	newEvalException(String message, Throwable throwable) Creates an exception which indicates that an error has occurred while executing a given function.
static RuntimeException	newEvalException(Throwable throwable) Creates an exception which indicates that an error has occurred while executing a given function.
static List<Member>	partiallySortMembers(Evaluator evaluator, List<Member> list, Calc exp, int limit, boolean desc) Partially sorts a list of Members by the value of an applied expression.
static List<List<Member>>	partiallySortTuples(Evaluator evaluator, TupleList list, Calc exp, int limit, boolean desc) Partially sorts a list of Tuples by the value of an applied expression.
protected static double	quartile(Evaluator evaluator, TupleList members, Calc exp, int range) Returns the member which lies upon a particular quartile according to a given expression.
static FunDef	resolveFunArgs(Validator validator, FunDef funDef, Exp[] args, Exp[] newArgs, String name, Syntax syntax) Validates the arguments to a function and resolves the function.
static TupleList	sortTuples(Evaluator evaluator, TupleIterable tupleIterable, TupleList tupleList, Calc exp, boolean desc, boolean brk, int arity) Sorts a list of Tuples by the value of an applied expression.
static <T> List<T>	stablePartialSort(List<T> list, Comparator<T> comp, int limit) Stable partial sort of a list.
static <T> List<T>	stablePartialSort(List<T> list, Comparator<T> comp, int limit, int algorithm) Stable partial sort of a list, using a specified algorithm.
static <T> List<T>	stablePartialSortArray(List<T> list, Comparator<T> comp, int limit) Partial sort an array by sorting it and returning the first limit elements.
static <T> List<T>	stablePartialSortJulian(List<T> list, Comparator<T> comp, int limit) Julian's algorithm for stable partial sort.
static <T> List<T>	stablePartialSortMarc(List<T> list, Comparator<T> comp, int limit) Marc's original algorithm for stable partial sort of a list.
static <T> List<T>	stablePartialSortPedro(List<T> list, Comparator<T> comp, int limit) Pedro's algorithm for stably sorting the top limit items in a list.
static Object	sum(Evaluator evaluator, TupleList members, Calc exp)
static double	sumDouble(Evaluator evaluator, TupleIterable iterable, Calc exp)
static double	sumDouble(Evaluator evaluator, TupleList members, Calc exp)
static boolean	tupleContainsNullMember(List<Member> tuple) Returns whether one of the members in a tuple is null.
static boolean	tupleContainsNullMember(Member[] tuple) Returns whether one of the members in a tuple is null.
static boolean	worthCaching(Exp exp) Returns whether an expression is worth wrapping in "Cache( ...

Methods inherited from class mondrian.olap.Util

addDatabaseTime, addLevelCalculatedMembers, append, appendArrays, areOccurencesEqual, assertPostcondition, assertPostcondition, assertPrecondition, assertPrecondition, assertTrue, assertTrue, badValue, binarySearch, bitSetBetween, camelToUpper, canCast, cancelStatement, caseSensitiveCompareName, cast, cast, castToIterable, checkCJResultLimit, close, commaList, compare, compareIntegers, compareKey, compareName, compileScript, convert, convert, convertOlap4jConnectStringToNativeMondrian, convertStackToString, copyOf, copyOf, copyOf, copyOf, copyOf, createExpr, createRandom, createRootRole, createSimpleValidator, createUdf, dbTimeMillis, deprecated, deprecated, digestMd5, digestSha256, equal, equalName, equals, equals, falseFunctor, filter, flatList, flatList, flatListCopy, generateUuidString, genericArray, getAnnotation, getClosestResource, getDimensionCube, getErrorMessage, getErrorMessage, getExecutorService, getFirstDescendantOnLevel, getMatchingCause, getMemberOrdinalInParent, getMembersToHierarchyMap, getMemoryInfo, getScheduledExecutorService, hash, hash, hashArray, identityFunctor, implode, intersect, isBlank, isEmpty, isNull, isSorted, isValidProperty, last, lookup, lookup, lookup, lookup, lookup, lookupCompound, lookupCompound, lookupHierarchyLevel, lookupHierarchyRootMember, lookupHierarchyRootMember, lookupProperty, makeBigDecimalFromDouble, makeFqName, makeFqName, makeFqName, matches, matches, mdxEncodeString, needToImplement, newElementNotFoundException, newError, newError, newIdentityHashSet, newIdentityHashSetFake, newInternal, newInternal, newTimer, nonDbTimeMillis, normalizeName, only, parseCommaList, parseConnectString, parseIdentifier, parseInterval, parseLocale, printMemory, printMemory, quoteForMdx, quoteForMdx, quoteJavaString, quoteMdxIdentifier, quoteMdxIdentifier, quoteMdxIdentifier, quoteMdxIdentifier, quotePattern, readFully, readFully, readURL, readURL, readURL, readVirtualFile, readVirtualFileAsString, replace, replace, replaceProperties, safeGet, singleQuoteString, singleQuoteString, sort, sort, threadLocalRemove, toMap, toNullValuesMap, toOlap4j, toOlap4j, toOlap4j, toURL, trueFunctor, unexpected, uniquify, unparse, unparse, wildcardToRegexp

Methods inherited from class org.eigenbase.xom.XOMUtil

addChild, addChild, addChildren, addElement, concatenate, createDefaultParser, discard, discard, discard, discard, discard, discard, discard, discard, discard, getFirstInstance, wrapperToXml

Methods inherited from class org.eigenbase.xom.XMLUtil

getFirstTagName, printAtt, printAtt, printAtt, printAtt, printPCDATA, printPCDATA, printPCDATA, quoteAtt, quoteAtt, quoteAtt, quoteAtt, quotePCDATA, stringEncodeXML, stringHasXMLSpecials

Methods inherited from class java.lang.Object

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

Field Detail

NullMember

public static final mondrian.olap.fun.FunUtil.NullMember NullMember

DoubleNull

public static final double DoubleNull

Special value which indicates that a double computation has returned the MDX null value. See DoubleCalc.

See Also:

Constant Field Values

DoubleEmpty

public static final double DoubleEmpty

Special value which indicates that a double computation has returned the MDX EMPTY value. See DoubleCalc.

See Also:

Constant Field Values

IntegerNull

public static final int IntegerNull

Special value which indicates that an int computation has returned the MDX null value. See IntegerCalc.

See Also:

Constant Field Values

BooleanNull

public static final boolean BooleanNull

Null value in three-valued boolean logic. Actually, a placeholder until we actually implement 3VL.

See Also:

Constant Field Values

Constructor Detail

FunUtil

public FunUtil()

Method Detail

newEvalException

public static RuntimeException newEvalException(FunDef funDef,
                                                String message)

Creates an exception which indicates that an error has occurred while executing a given function.

Parameters:

funDef - Function being executed

message - Explanatory message

Returns:

Exception that can be used as a cell result

newEvalException

public static RuntimeException newEvalException(Throwable throwable)

Creates an exception which indicates that an error has occurred while executing a given function.

Parameters:

throwable - Exception

Returns:

Exception that can be used as a cell result

newEvalException

public static RuntimeException newEvalException(String message,
                                                Throwable throwable)

Creates an exception which indicates that an error has occurred while executing a given function.

Parameters:

message - Explanatory message

throwable - Exception

Returns:

Exception that can be used as a cell result

checkIterListResultStyles

public static void checkIterListResultStyles(Calc calc)

checkListResultStyles

public static void checkListResultStyles(Calc calc)

getDimensionDefaultHierarchy

public static Hierarchy getDimensionDefaultHierarchy(Dimension dimension)

Returns the default hierarchy of a dimension, or null if there is no default.

Parameters:

dimension - Dimension

Returns:

Default hierarchy, or null

See Also:

MondrianResource.CannotImplicitlyConvertDimensionToHierarchy

sortTuples

public static TupleList sortTuples(Evaluator evaluator,
                                   TupleIterable tupleIterable,
                                   TupleList tupleList,
                                   Calc exp,
                                   boolean desc,
                                   boolean brk,
                                   int arity)

Sorts a list of Tuples by the value of an applied expression. Stable sort.

Helper function for MDX functions TopCount, TopSum, TopPercent, BottomCount, BottomSum, BottomPercent, but not the MDX function Order.

NOTE: This function does not preserve the contents of the validator.

If you specify tupleList, the list is sorted in place, and tupleList is returned.

Parameters:

evaluator - Evaluator

tupleIterable - Iterator over tuples

tupleList - List of tuples, if known, otherwise null

exp - Expression to sort on

desc - Whether to sort descending

brk - Whether to break

arity - Number of members in each tuple

Returns:

sorted list (never null)

partiallySortMembers

public static List<Member> partiallySortMembers(Evaluator evaluator,
                                                List<Member> list,
                                                Calc exp,
                                                int limit,
                                                boolean desc)

Partially sorts a list of Members by the value of an applied expression.

Avoids sorting the whole list, finds only the ntop (or bottom) valued Members, and returns them as a new List. Helper function for MDX functions TopCount and BottomCount.

NOTE: Does not preserve the contents of the validator.

Parameters:

list - a list of members

exp - a Calc applied to each member to find its sort-key

evaluator - Evaluator

limit - maximum count of members to return.

desc - true to sort descending (and find TopCount), false to sort ascending (and find BottomCount).

Returns:

the top or bottom members, as a new list.

partiallySortTuples

public static List<List<Member>> partiallySortTuples(Evaluator evaluator,
                                                     TupleList list,
                                                     Calc exp,
                                                     int limit,
                                                     boolean desc)

Partially sorts a list of Tuples by the value of an applied expression.

Avoids sorting the whole list, finds only the n top (or bottom) valued Tuples, and returns them as a new List. Helper function for MDX functions TopCount and BottomCount.

NOTE: Does not preserve the contents of the validator. The returned list is immutable.

Parameters:

evaluator - Evaluator

list - a list of tuples

exp - a Calc applied to each tuple to find its sort-key

limit - maximum count of tuples to return.

desc - true to sort descending (and find TopCount), false to sort ascending (and find BottomCount).

Returns:

the top or bottom tuples, as a new list.

hierarchizeMemberList

public static void hierarchizeMemberList(List<Member> memberList,
                                         boolean post)

Sorts a list of members into hierarchical order. The members must belong to the same dimension.

Parameters:

memberList - List of members

post - Whether to sort in post order; if false, sorts in pre order

See Also:

hierarchizeTupleList(mondrian.calc.TupleList, boolean)

hierarchizeTupleList

public static TupleList hierarchizeTupleList(TupleList tupleList,
                                             boolean post)

Sorts a list of tuples into hierarchical order.

Parameters:

tupleList - List of tuples

post - Whether to sort in post order; if false, sorts in pre order

See Also:

hierarchizeMemberList(java.util.List, boolean)

compareValues

public static int compareValues(double d1,
                                double d2)

Compares double-precision values according to MDX semantics.

MDX requires a total order:

-inf < NULL < ... < -1 < ... < 0 < ... < NaN < +inf

but this is different than Java semantics, specifically with regard to Double.NaN.

compareValues

public static int compareValues(Object value0,
                                Object value1)

Compares two cell values.

Nulls compare last, exceptions (including the object which indicates the the cell is not in the cache yet) next, then numbers and strings are compared by value.

Parameters:

value0 - First cell value

value1 - Second cell value

Returns:

-1, 0, or 1, depending upon whether first cell value is less than, equal to, or greater than the second

decodeSyntacticType

public static Syntax decodeSyntacticType(String flags)

Decodes the syntactic type of an operator.

Parameters:

flags - A encoded string which represents an operator signature, as used by the flags parameter used to construct a FunDefBase.

Returns:

A Syntax

decodeReturnCategory

public static int decodeReturnCategory(String flags)

Decodes the signature of a function into a category code which describes the return type of the operator.

For example, decodeReturnType("fnx") returns Category.Numeric, indicating this function has a numeric return value.

Parameters:

flags - The signature of an operator, as used by the flags parameter used to construct a FunDefBase.

Returns:

An array Category codes.

decodeCategory

public static int decodeCategory(String flags,
                                 int offset)

Decodes the offsetth character of an encoded method signature into a type category.

The codes are:

a	Category.Array
d	Category.Dimension
h	Category.Hierarchy
l	Category.Level
b	Category.Logical
m	Category.Member
N	Constant Category.Numeric
n	Category.Numeric
x	Category.Set
#	Constant Category.String
S	Category.String
t	Category.Tuple
v	Category.Value
y	Category.Symbol

Parameters:

flags - Encoded signature string

offset - 0-based offset of character within string

Returns:

A Category

decodeParameterCategories

public static int[] decodeParameterCategories(String flags)

Decodes a string of parameter types into an array of type codes.

Each character is decoded using decodeCategory(String, int). For example, decodeParameterTypes("nx") returns {Category.Numeric, Category.Set}.

Parameters:

flags - The signature of an operator, as used by the flags parameter used to construct a FunDefBase.

Returns:

An array Category codes.

box

public static Double box(double d)

Converts a double (primitive) value to a Double. DoubleNull becomes null.

box

public static Integer box(int n)

Converts an int (primitive) value to an Integer. IntegerNull becomes null.

quartile

protected static double quartile(Evaluator evaluator,
                                 TupleList members,
                                 Calc exp,
                                 int range)

Returns the member which lies upon a particular quartile according to a given expression.

Parameters:

evaluator - Evaluator

members - List of members

exp - Expression to rank members

range - Quartile (1, 2 or 3)

min

public static Object min(Evaluator evaluator,
                         TupleList members,
                         Calc calc)

max

public static Object max(Evaluator evaluator,
                         TupleList members,
                         Calc exp)

avg

public static Object avg(Evaluator evaluator,
                         TupleList members,
                         Calc calc)

sum

public static Object sum(Evaluator evaluator,
                         TupleList members,
                         Calc exp)

sumDouble

public static double sumDouble(Evaluator evaluator,
                               TupleList members,
                               Calc exp)

sumDouble

public static double sumDouble(Evaluator evaluator,
                               TupleIterable iterable,
                               Calc exp)

count

public static int count(Evaluator evaluator,
                        TupleIterable iterable,
                        boolean includeEmpty)

compareHierarchically

public static int compareHierarchically(Member m1,
                                        Member m2,
                                        boolean post)

Compares a pair of members according to their positions in a prefix-order (or postfix-order, if post is true) walk over a hierarchy.

Parameters:

m1 - First member

m2 - Second member

post - Whether to sortMembers in postfix order. If true, a parent will sortMembers immediately after its last child. If false, a parent will sortMembers immediately before its first child.

Returns:

-1 if m1 collates before m2, 0 if m1 equals m2, 1 if m1 collates after m2

compareSiblingMembers

public static int compareSiblingMembers(Member m1,
                                        Member m2)

Compares two members which are known to have the same parent. First, compare by ordinal. This is only valid now we know they're siblings, because ordinals are only unique within a parent. If the dimension does not use ordinals, both ordinals will be -1.

If the ordinals do not differ, compare using regular member comparison.

Parameters:

m1 - First member

m2 - Second member

Returns:

-1 if m1 collates less than m2, 1 if m1 collates after m2, 0 if m1 == m2.

tupleContainsNullMember

public static boolean tupleContainsNullMember(Member[] tuple)

Returns whether one of the members in a tuple is null.

tupleContainsNullMember

public static boolean tupleContainsNullMember(List<Member> tuple)

Returns whether one of the members in a tuple is null.

makeNullTuple

public static Member[] makeNullTuple(TupleType tupleType)

resolveFunArgs

public static FunDef resolveFunArgs(Validator validator,
                                    FunDef funDef,
                                    Exp[] args,
                                    Exp[] newArgs,
                                    String name,
                                    Syntax syntax)

Validates the arguments to a function and resolves the function.

Parameters:

validator - Validator used to validate function arguments and resolve the function

funDef - Function definition, or null to deduce definition from name, syntax and argument types

args - Arguments to the function

newArgs - Output parameter for the resolved arguments

name - Function name

syntax - Syntax style used to invoke function

Returns:

resolved function definition

getNonEmptyMemberChildren

public static List<Member> getNonEmptyMemberChildren(Evaluator evaluator,
                                                     Member member)

getNonEmptyMemberChildrenWithDetails

public static Map<Member,Access> getNonEmptyMemberChildrenWithDetails(Evaluator evaluator,
                                                                      Member member)

stablePartialSort

public static <T> List<T> stablePartialSort(List<T> list,
                                            Comparator<T> comp,
                                            int limit)

Stable partial sort of a list. Returns the desired head of the list.

stablePartialSort

public static <T> List<T> stablePartialSort(List<T> list,
                                            Comparator<T> comp,
                                            int limit,
                                            int algorithm)

Stable partial sort of a list, using a specified algorithm.

stablePartialSortArray

public static <T> List<T> stablePartialSortArray(List<T> list,
                                                 Comparator<T> comp,
                                                 int limit)

Partial sort an array by sorting it and returning the first limit elements. Fastest approach if limit is a significant fraction of the list.

stablePartialSortMarc

public static <T> List<T> stablePartialSortMarc(List<T> list,
                                                Comparator<T> comp,
                                                int limit)

Marc's original algorithm for stable partial sort of a list. Now superseded by stablePartialSortJulian(java.util.List<T>, java.util.Comparator<T>, int).

stablePartialSortPedro

public static <T> List<T> stablePartialSortPedro(List<T> list,
                                                 Comparator<T> comp,
                                                 int limit)

Pedro's algorithm for stably sorting the top limit items in a list.

stablePartialSortJulian

public static <T> List<T> stablePartialSortJulian(List<T> list,
                                                  Comparator<T> comp,
                                                  int limit)

Julian's algorithm for stable partial sort. Improves Pedro's algorithm by using a heap (priority queue) for the top limit items seen. The items on the priority queue have an ordinal field, so the queue can be used to generate a list of stably sorted items. (Heap sort is not normally stable.)

Type Parameters:

T - Element type

Parameters:

list - List to sort

comp - Comparator

limit - Maximum number of items to return

Returns:

Sorted list, containing at most limit items

worthCaching

public static boolean worthCaching(Exp exp)

Returns whether an expression is worth wrapping in "Cache( ... )".

Parameters:

exp - Expression

Returns:

Whether worth caching