Transcript Object-Relational Features - Department of Computer and

Department of Computer and Information Science,
School of Science, IUPUI
Object-Relational Features
Dale Roberts, Lecturer
Computer Science, IUPUI
E-mail: [email protected]
Dale Roberts
1
Object-Relational Databases
Types, Object Views, Methods
Nested Tables and Varying Arrays
Large Objects
Advanced OO Concepts
Dale Roberts
2
Types
Abstract datatypes groups related columns/data .
create type ADDRESS_TY as object
( Street VARCHAR2(50),
City
VARCHAR2(25),
State
CHAR(2),
Zip
NUMBER);
Abstract datatypes can be nested .
security – must have EXECUTE privilege to use a type.
set describe depth – see attributes within an object.
Dale Roberts
3
Types: indexing, selecting, manipulating
Create type PERSON_TY (
Name VARCHAR2(25),
Address ADDRESS_TY);
Create table CUSTOMER (
Cust_Id NUMBER,
Person PERSON_TY );
Select:
select c.Name, c.Person.Address.City
from CUSTOMER c
where c.Person.Address.City like ‘I%’;
Indexing:
create index IDX_CUST on CUSTOMER
(Person.Address.City);
Insert:
insert into CUSTOMER values
(1, PERSON_TY(‘xxxx’,
ADDRESS_TY(‘aaaa’, ‘bbb’, ccc’)));
Dale Roberts
4
Object Views
create table CUSTOMER
(Cust_id
NUMBER PRIMARY KEY,
Name
VARCHAR2(25),
Street
VARCHAR2(50),
City
VARCHAR2(25),
State
CHAR(2),
Zip
NUMBER);
create type ADDRESS_TY as object
( Street
VARCHAR2(50),
City
VARCHAR2(25),
State
CHAR(2),
Zip
NUMBER);

Inserting into original table:
insert into CUSTOMER values
( 123,
‘name’,
‘Street’, ‘City’,
‘ST’, 99999 );

create type PERSON_TY (
Name
VARCHAR2(25),
Address ADDRESS_TY);
create or replace view CUSTOMER_OV
(Cust_id, person) as
select Cust_id,
PERSON_TY (Name,
ADDRESS_TY(street, city, state, zip))
from CUSTOMER;
Dale Roberts
Inserting into object view:
insert into CUSTOMER_OV values
( 123,
PERSON_TY(‘name’,
ADDRESS_TY(
‘Street’, ‘City’,
‘ST’, 99999 )));
5
INSTEAD OF Triggers
Views usually cannot use standard table-based PL/SQL Triggers,
but they can use INSTEAD OF triggers. Works with either normal views or
object views.
create trigger AUTHOR_PUBLISHER_UPDATE
instead of UPDATE on AUTHOR_PUBLISHER
for each row
begin
…
end;
Dale Roberts
6
Methods
Methods allow types to have “behavior”.
Methods are declared inside CREATE TYPE,
they are defined with CREATE TYPE BODY.
create type PERSON_TY (
Name
VARCHAR2(25),
Address ADDRESS_TY,
Birthday DATE,
member function AGE (BirthDate IN DATE) return NUMBER);
create or replace type body PERSON_TY as
member function AGE (Birthdate DATE) return NUMBER is
begin
return ROUND(sysdate – Birthdate);
end;
end;
Dale Roberts
7
Collections: Varying Arrays

Definition – a varying array is a set of objects, each of same type, seen as a
column within a table. The size is limited when created.

Creating a Varying Array
1. Create type TOOL_TY as object (…);
2. Create type TOOLS_VA as varray(5) of VARCHAR2(25);
3. Create table TOOL_BORROWER (…, Tools TOOLS_VA);

Describing – requires queries on data dictionary to show varray structure.

Selecting – requires PL/SQL, see p. 571-572

Inserting –
insert into TOOL_BORROWER values (‘name’,
TOOLS_VA(‘tool1’, ‘tool2’, ‘tool3’));

Storage – stored in-line, ie. within table data segment.
Dale Roberts
8
Collections: Nested Tables

Definition – a nested table is a table within a table.
Their size is not limited as with varying arrays.

Creating a Nested Table
1. Create type TOOL_TY as object (…);
2. Create type TOOLS_NT as table of TOOL_TY;
3. Create table TOOL_BORROWER (…,
nested table Tools TOOLS_NT);

Describing – set describe depth will allow the describe command to show nested tables.

Selecting – requires the TABLE() function (previous to 9i use THE() function).
This “flatten’s” or de-normalizes the nested table’s relationship.
Select BreederName, N.Name, N.birthdate
from BREEDER, TABLE(breeder.Animals) N

Inserting –

Storage – Nested tables can be stored in or out-of-line. Recommend they be stored
out-of-line with a storage clause to improves query performance.
insert into TABLE(select Animals
from BREEDER where Breedername = ‘x’)
values (ANIMAL_TY (‘x’,’y’,’z’));
Dale Roberts
9
Large Object Datatypes

BLOB – binary field up to 4Gb.

CLOB – character field up to 4Gb.

NCLOB – CLOB that supports MBCS (multi-byte character set).

BFILE – read-only binary data stored in a file outside of the database.
Dale Roberts
10
LOB Storage, Initialization

LOB Storage – BLOB and (N)CLOB columns are stored out-of-line.
They require the CREATE TABLE statement include a LOB
STORAGE clause to tell where the LOB segment is stored. BFILE
columns are by their nature stored out-of-line.

LOBs use at least 1 entire data block per row in the LOB segment.
Normal DATA segments store multiple rows per block (up to 100% PCTFREE).

Initializing LOBs – use EMPTY_BLOB or EMPTY_CLOB instead of
NULL to initialize a LOB.
• NULL will create a LOB with no data.
• EMPTY_BLOB/CLOB will create an empty “LOB locater” (pointer)
without allocating the empty block.
Dale Roberts
11
DBMS_LOB

LOB data is manipulated with the DBMS_LOB package.

For all LOBs:
•
•
•
•
•
•
•
•
•
•
•
•
•
•
APPEND
COMPARE
COPY
ERASE
GETCHUNKSIZE
GETLENGTH
INSTR
ISOPEN
OPEN
READ
SUBSTR
TRIM
WRITE
WRITEAPPEND


BFILE specific functions:
•
•
•
•
•
•
•
FILEOPEN
FILECLOSE
FILECLOSEALL
FILEEXISTS
FILEGETNAME
FILEISOPEN
LOADFROM FILE
Temporary LOB functions:
•
•
•
CREATETEMPORARY
FREETEMPORARY
ISTEMPORARY
Dale Roberts
12
BFILE

Read-only binary data stored in a file outside of the database.
CREATE DIRECTORY PROPOSAL_DIR AS ‘C:\My Documents’;
Insert into thetable ( col1, bfilecol ) values
( ‘value1’, BFILENAME(‘proposal_dir’, ‘myfile.doc’);

Deleting LOBs – internal LOBs when deleted have both their data and
LOB locater deleted. Deleting a BFILE only deletes the locater, the file
is read only and thus not changed.
Dale Roberts
13
row objects vs. column objects

Row objects

Column objects
Dale Roberts
14
REF, DEREF, VALUE

REF

DEREF

VALUE
Dale Roberts
15
Acknowledgements
Loney, Oracle Database 10g The Complete Reference
4/2/2016
Dale Roberts
16