filter

Value Members

1. object Bounds extends Serializable

   

2. object FilterHelper

   

3. object FilterProperties

   

4. object FilterValues extends Serializable

   

5. object GeometryProcessing extends GeometryProcessing with LazyLogging

   

6. object RichTransform

   

7. val WildcardMultiChar: Char

   

8. val WildcardSingleChar: Char

   

9. val WildcardSuffix: String

   

10. val Wildcards: Seq[Char]

    

11. def andFilters(filters: Seq[Filter])(implicit ff: FilterFactory = ff): Filter

    

12. def andOption(filters: Seq[Filter])(implicit ff: FilterFactory = ff): Option[Filter]

    

13. def checkOrder(one: Expression, two: Expression): Option[AttributeExpression]

    

    Checks the order of properties and literals in the expression

    Checks the order of properties and literals in the expression

    one

    first expression

    two

    second expression

    returns

    (prop, literal, whether the order was flipped)

14. def checkOrderUnsafe(one: Expression, two: Expression): AttributeExpression

    

    Checks the order of properties and literals in the expression - if the expression does not contain a property and a literal, throws an exception.

    Checks the order of properties and literals in the expression - if the expression does not contain a property and a literal, throws an exception.

    one

    first expression

    two

    second expression

    returns

    (prop, literal, whether the order was flipped)

15. def decomposeAnd(f: Filter): Seq[Filter]

    

16. def decomposeBinary(f: Filter): Seq[Filter]

    

17. def decomposeOr(f: Filter): Seq[Filter]

    

18. package expression

    

19. package factory

    

20. implicit val ff: FilterFactory2

    

21. def filterToString(filter: Option[Filter]): String

    

22. def filterToString(filter: Filter): String

    

23. def filtersToString(filters: Seq[Filter]): String

    

24. package function

    

25. def getAttributeProperty(f: Filter): Option[String]

    

26. package index

    

27. def isIdFilter(f: Filter): Boolean

    

28. def isPrimarySpatialFilter(filter: Filter, sft: SimpleFeatureType): Boolean

    

29. def isPrimaryTemporalFilter(f: Filter, sft: SimpleFeatureType): Boolean

    

30. def isSpatialFilter(f: Filter): Boolean

    

31. def isTemporalFilter(f: Filter, dtg: String): Boolean

    

32. def likeEligible(filter: PropertyIsLike): Boolean

    

    Like queries we can handle with our current index

    Like queries we can handle with our current index

    filter

    filter

33. def logicDistributionCNF(x: Filter): List[List[Filter]]

    
    returns

    A List[ List[Filter] ] where the inner List of Filters are to be joined by Ors and the outer list combined by Ands.

34. def mergeFilters(f1: Filter, f2: Filter): Filter

    

35. def orFilters(filters: Seq[Filter])(implicit ff: FilterFactory = ff): Filter

    

36. def orOption(filters: Seq[Filter])(implicit ff: FilterFactory = ff): Option[Filter]

    

37. def partitionID(filter: Filter): (Seq[Filter], Seq[Filter])

    

38. def partitionPrimarySpatials(filters: Seq[Filter], sft: SimpleFeatureType): PartionedFilter

    

39. def partitionPrimarySpatials(filter: Filter, sft: SimpleFeatureType): PartionedFilter

    

40. def partitionPrimaryTemporals(filters: Seq[Filter], sft: SimpleFeatureType): PartionedFilter

    

41. def partitionSubFilters(filter: Filter, filterFilter: (Filter) ⇒ Boolean): PartionedFilter

    

42. def rewriteFilterInCNF(filter: Filter)(implicit ff: FilterFactory): Filter

    

    This function rewrites a org.opengis.filter.Filter in terms of a top-level AND with children filters which 1) do not contain further ANDs, (i.e., ANDs bubble up) 2) only contain at most one OR which is at the top of their 'tree'

    This function rewrites a org.opengis.filter.Filter in terms of a top-level AND with children filters which 1) do not contain further ANDs, (i.e., ANDs bubble up) 2) only contain at most one OR which is at the top of their 'tree'

    Note that this further implies that NOTs have been 'pushed down' and do have not have ANDs nor ORs as children.

    In boolean logic, this form is called conjunctive normal form (CNF).

    The main use case for this function is to aid in splitting filters between a combination of a GeoMesa data store and some other data store. This is done with the AndSplittingFilter class. In the examples below, anything with "XAttr" is assumed to be a filter that CANNOT be answered through GeoMesa. In having a filter split on the AND, the portion of the filter that GeoMesa CAN answer will be applied in GeoMesa, returning a result set, and then the portion that GeoMesa CANNOT answer will be applied on that result set.

    Examples:

    1. ( (GmAttr ILIKE 'test') OR (date BETWEEN '2014-01-01T10:30:00.000Z' AND '2014-01-02T10:30:00.000Z') ) AND (XAttr ILIKE = 'example')

    Converting to CNF will allow easily splitting the filter on the AND into two children

    • one child is the "GmAttr" and "date" filters that can be answered with GeoMesa
    • one child is the "XAttr" filter that cannot be answered by GeoMesa

    In this case, the GeoMesa child filter will be processed first, and then the "XAttr" filter will be processed on the GeoMesa result set to return a subset of the GeoMesa results.

    2. (GmAttr ILIKE 'test') AND ( (date BETWEEN '2014-01-01T10:30:00.000Z' AND '2014-01-02T10:30:00.000Z') OR (XAttr1 ILIKE = 'example1') ) AND (XAttr2 ILIKE = 'example2')

    Converting to CNF still allows easily splitting the filter on the AND into three children

    • one child is the "GmAttr" filter
    • one child is the "date" OR "XAttr1" filter
    • one child is the "XAttr2" filter

    In this case, the "GmAttr" child will be processed first, returning a result set from GeoMesa called RS1. Then, RS1 will be further filtered with the "date" predicate that can be handled by GeoMesa, returning a subset of RS1 called SS1. The additional filter which cannot be answered by GeoMesa, "XAttr1," will be applied to RS1 and return subset SS2. Finally, the final child, the "XAttr2" filter, which cannot be answered by GeoMesa, will be applied to both SS1 and SS2 to return SS3, a JOIN of SS1+SS2 filtered with "XAttr2."

    3. (GmAttr ILIKE 'test') OR (XAttr ILIKE = 'example')

    This is the worst-case-scenario for a query that is answered through two data stores, both GeoMesa and some other store.

    CNF converts this to:

    • one child of "GmAttr" OR "XAttr"

    In this case, the "GmAttr" will return a result set, RS1. The reason this is the worst-case-scenario is because, to answer the "XAttr" portion of the query (which cannot be answered by GeoMesa), a "Filter.INCLUDE" A.K.A a full table scan (on Accumulo) A.K.A. every record in GeoMesa is necessary to find the results that satisfy the "XAttr" portion of the query. This will product result set RS2. The returned results will be a JOIN of RS1+RS2.

    filter

    An arbitrary filter.

    returns

    A filter in CNF (described above).

43. def rewriteFilterInDNF(filter: Filter)(implicit ff: FilterFactory): Filter

    

    This function rewrites a org.opengis.filter.Filter in terms of a top-level OR with children filters which 1) do not contain further ORs, (i.e., ORs bubble up) 2) only contain at most one AND which is at the top of their 'tree'

    This function rewrites a org.opengis.filter.Filter in terms of a top-level OR with children filters which 1) do not contain further ORs, (i.e., ORs bubble up) 2) only contain at most one AND which is at the top of their 'tree'

    Note that this further implies that NOTs have been 'pushed down' and do have not have ANDs nor ORs as children.

    In boolean logic, this form is called disjunctive normal form (DNF).

    filter

    An arbitrary filter.

    returns

    A filter in DNF (described above).

44. package visitor