OraGIST How to Make User-Defined Indexing Become Usable and Useful - - PowerPoint PPT Presentation

▶
oragist how to make user defined indexing become usable
SMART_READER_LITE
LIVE PREVIEW

OraGIST How to Make User-Defined Indexing Become Usable and Useful - - PowerPoint PPT Presentation

Dec 29, 2022 133 likes •312 views

1 OraGIST How to Make User-Defined Indexing Become Usable and Useful Carsten Kleiner, Udo Lipeck Universit at Hannover BTW 2003, 27.02.2003 2 The Situation (+) Object-relational DBMSs are extensible by: user-defined datatypes

slide-1
SLIDE 1

1

OraGIST – How to Make User-Defined Indexing Become Usable and Useful

Carsten Kleiner, Udo Lipeck Universit¨ at Hannover

BTW 2003, 27.02.2003

slide-2
SLIDE 2

2

The Situation (+)

Object-relational DBMSs are extensible by:

  • user-defined datatypes (UDT)
  • together with corresponding query operators
  • user-defined indexing
  • together with corresponding user-defined
  • ptimizer tuning

⇒ ”data blades”, ”cartridges”, ”extenders”, ...

slide-3
SLIDE 3

3

The Situation (−)

But user-defined indexing

  • needs expensive implementation
  • needs experimental selection for new query scenarios
  • is restricted:

a user-defined index supports only single operators, no combinations

  • assumes that an operator is supported by at most
  • ne index
slide-4
SLIDE 4

4

Our Proposal

  • Use combined datatypes and operators to allow

index support for attribute combinations

  • Provide medium-dimensional index structures
  • Apply a generic indexing framework

⋆ based on generalized search trees (GiST) ⋆ easily specializable into indexes for particular UDTs and operators ⋆ coupled with DB storage and index definition ⇒ prototypical tool OraGiST (GiST for Oracle)

slide-5
SLIDE 5

5

Generalized Search Trees (GiST) — Class Hierarchy

GiST

#insert(e:GiSTEntry,level:int) #chooseSubtree(e:GiSTEntry,level:int) #split(n:GiSTNode,e:GiSTEntry) #adjustKeys(n:GiSTNode) #delete(e:GiSTEntry)

OrderedGiST

#findMin(q:predicate) #next(q:predicate,e:GiSTEntry)

UnorderedGiST

#search(q:predicate)

BTreeGiST

#consistent(e:GiSTEntry,q:predicate) #union(l:ListOfGiSTEntry) #penalty(e1:GiSTEntry,e2:GiSTEntry) #pickSplit(l:ListOfGiSTEntry) #compare(e1:GiSTEntry,e2:GiSTEntry)

RTreeGiST

#consistent(e:GiSTEntry,q:predicate) #union(l:ListOfGiSTEntry) #penalty(e1:GiSTEntry,e2:GiSTEntry) #pickSplit(l:ListOfGiSTEntry)

RStarTreeGiST

#penalty(e1:GiSTEntry,e2:GiSTEntry) #pickSplit(l:ListOfGiSTEntry)

SSTreeGiST

#penalty(e1:GiSTEntry,e2:GiSTEntry)

RSSTreeGiST

#penalty(e1:GiSTEntry,e2:GiSTEntry)

slide-6
SLIDE 6

6

Index Usage

CREATE INDEXTYPE polygonRTree FOR overlaps(polygon,polygon) USING overlaps fct; CREATE TABLE county (id NUMBER, population NUMBER, ... shape polygon, ...); CREATE INDEX idx geoCounty ON county(shape) INDEXTYPE IS polygonRTree PARAMETERS(...); SELECT * FROM county WHERE overlaps(shape,polygon(rectangle(0,0,20,10)))) = true;

slide-7
SLIDE 7

7

Extensible Indexing in ORDBS

  • Sequence of index operations over index lifetime (e.g.):

ODCIIndexCreate ODCIIndexInsert ODCIIndexDrop for each tuple

  • Sequence of index operations in query execution:

ODCIIndexFetch ODCIIndexClose

delete structures set up structures

ODCIIndexStart Scan Context

read and modify

until all tuples fetched

slide-8
SLIDE 8

8

OraGiST — Architecture and Functionality

GiST GiSTExtension GiSTEntry

1 n

GiSTIndexFile DBSExtensibleIndexing UserDefinedIndexStructure DBSUserDefinedObject

1 n

DBSIndexTable

Oracle ORDBS

1 1

libgist library

1 1

OraGiST

OraGiST Library OraGiST Toolbox

initiates calls is stored in 1 n

OraGiSTExtension

+getExtension() +getQuery() +needVerify()

TypeMap

+approximate()

slide-9
SLIDE 9

9

OraGiST — Architecture and Functionality

GiST RTree RTreeEntry

1 n

GiSTIndexFile OracleExtensibleIndexing UserDefinedRTree GeometryObject

1 n

OracleIndexTable

Oracle ORDBS

1 1

libgist library

1 1

OraGiST

getExtension getQuery approximate

slide-10
SLIDE 10

10

Case Study — Spatial/Thematic Analysis

  • Typical query: Find all counties overlapping a given

window where the median rent is below ... and the population is higher than ...

  • Definition of combined datatypes and operators:

CREATE TYPE integerGeometry AS OBJECT (shape OGCGeometry, theme INTEGER); CREATE TYPE twoIntegerGeometry AS OBJECT (shape OGCGeometry, theme1,theme2 INTEGER); CREATE OPERATOR [two]BetweenOverlaps ... ; CREATE TABLE county OF [two]IntegerGeometry;

slide-11
SLIDE 11

11

Performance Evaluation on 2D Spatial Data

10 20 30 40 50 60 5 10 15 20 25 30 35 40 response time selectivity in % Oracle R-Tree Oracle Quadtree User-Def-R*-Tree

slide-12
SLIDE 12

12

... on 2D spatial data + one thematic dimension

10 20 30 40 50 60 70 80 2 4 6 8 10 12 14 response time selectivity in % Oracle R-Tree User-Def-3D-RSS-Tree User-Def-3D-R*-Tree Oracle Quadtree

slide-13
SLIDE 13

13

... on 2D spatial data + two thematic dimensions

10 20 30 40 50 60 70 80 5 10 15 20 response time selectivity in % User-Def-4D-RSS-Tree User-Def-4D-R*-Tree User-Def-3D-RSS-Tree User-Def-2D-RSS-Tree

slide-14
SLIDE 14

14

Conclusions

  • R∗- and RSS-trees can be recommended as (medium-

dimensional) index types for combined query scenarios.

  • Combined scenarios occur often in spatio-temporal-

thematic DBS; think of, e.g., ”validtimeInteger” !

  • For new UDTs/operators,

for combined UDTs/operators, and for experimental index selection, a framework for adaptable indexing in ORDBS is required.

slide-15
SLIDE 15

15

Conclusions (cont.)

  • OraGiST is such an indexing framework:

⋆ extensible library + toolbox ⋆ coupling the GiST-family with an ORDBMS ⋆ for adoption of existing index types ⋆ for fast development of new index types by object-oriented specialization

slide-16
SLIDE 16

16

Future Questions

  • How to improve behaviour on ’complicated’ objects ?
  • How to include user-defined cost/selectivity estimation ?
  • How to generate and hide the combined operators ?