Materialized Views in Oracle

A materialized view, or snapshot as they were previously known, is a table segment whose contents are periodically refreshed based on a query, either against a local or remote table. Using materialized views against remote tables is the simplest way to achieve replication of data between sites. The example code in this article assumes DB1 is the master instance and DB2 is the materialized view site.

Basic Syntax

The full syntax description for the CREATE MATERIALIZED VIEW command is available in the documentation. Here we will only concern ourselves with the basics.

CREATE MATERIALIZED VIEW view-name
BUILD [IMMEDIATE | DEFERRED]
REFRESH [FAST | COMPLETE | FORCE ]
ON [COMMIT | DEMAND ]
[[ENABLE | DISABLE] QUERY REWRITE]
[ON PREBUILT TABLE]
AS
SELECT ...;

The BUILD clause options are shown below.

IMMEDIATE : The materialized view is populated immediately.
DEFERRED : The materialized view is populated on the first requested refresh.

The following refresh types are available.

FAST : A fast refresh is attempted. If materialized view logs are not present against the source tables in advance, the creation fails.
COMPLETE : The table segment supporting the materialized view is truncated and repopulated completely using the associated query.
FORCE : A fast refresh is attempted. If one is not possible a complete refresh is performed.

A refresh can be triggered in one of two ways.

ON COMMIT : The refresh is triggered by a committed data change in one of the dependent tables.
ON DEMAND : The refresh is initiated by a manual request or a scheduled task.

The QUERY REWRITE clause tells the optimizer if the materialized view should be consider for query rewrite operations. An example of the query rewrite functionality is shown below.
The ON PREBUILT TABLE clause tells the database to use an existing table segment, which must have the same name as the materialized view and support the same column structure as the query.

Check Privileges

Check the user who will own the materialized views has the correct privileges. At minimum they will require the CREATE MATERIALIZED VIEW privilege. If they are creating materialized views using database links, you may want to grant them CREATE DATABASE LINK privilege also.

CONNECT sys@db2

GRANT CREATE MATERIALIZED VIEW TO scott;
GRANT CREATE DATABASE LINK TO scott;

Create Materialized View

Connect to the materialized view owner and create the database link and the materialized view itself.

CONNECT scott/tiger@db2

CREATE DATABASE LINK DB1.WORLD CONNECT TO scott IDENTIFIED BY tiger USING 'DB1.WORLD';

CREATE MATERIALIZED VIEW emp_mv
BUILD IMMEDIATE 
REFRESH FORCE
ON DEMAND
AS
SELECT * FROM emp@db1.world;

Alternatively, we could have used a prebuilt table, as shown below.

-- Create the tale first. This could be populated
-- using an export/import.
CREATE TABLE emp_mv AS
SELECT * FROM emp@db1.world;

-- Build the materialized view using the existing table segment.
CREATE MATERIALIZED VIEW emp_mv
REFRESH FORCE
ON DEMAND
ON PREBUILT TABLE
AS
SELECT * FROM emp@db1.world;

Remember to gather stats after building the materialized view.

BEGIN
  DBMS_STATS.gather_table_stats(
    ownname => 'SCOTT',
    tabname => 'EMP_MV');
END;
/

Create Materialized View Logs

Since a complete refresh involves truncating the materialized view segment and re-populating it using the related query, it can be quite time consuming and involve a considerable amount of network traffic when performed against a remote table. To reduce the replication costs, materialized view logs can be created to capture all changes to the base table since the last refresh. This information allows a fast refresh, which only needs to apply the changes rather than a complete refresh of the materialized view.
To take advantage of the of the fast refresh, connect to the master instance and create the materialized view log.

CONNECT scott/tiger@db1

CREATE MATERIALIZED VIEW LOG ON scott.emp
TABLESPACE users
WITH PRIMARY KEY
INCLUDING NEW VALUES;

Refresh Materialized Views

If a materialized view is configured to refresh on commit, you should never need to manually refresh it, unless a rebuild is necessary. Remember, refreshing on commit is a very intensive operation for volatile base tables. It makes sense to use fast refreshes where possible.
For on demand refreshes, you can choose to manually refresh the materialized view or refresh it as part of a refresh group.
The following code creates a refresh group defined to refresh every minute and assigns a materialized view to it.

BEGIN
   DBMS_REFRESH.make(
     name                 => 'SCOTT.MINUTE_REFRESH',
     list                 => '',
     next_date            => SYSDATE,
     interval             => '/*1:Mins*/ SYSDATE + 1/(60*24)',
     implicit_destroy     => FALSE,
     lax                  => FALSE,
     job                  => 0,
     rollback_seg         => NULL,
     push_deferred_rpc    => TRUE,
     refresh_after_errors => TRUE,
     purge_option         => NULL,
     parallelism          => NULL,
     heap_size            => NULL);
END;
/

BEGIN
   DBMS_REFRESH.add(
     name => 'SCOTT.MINUTE_REFRESH',
     list => 'SCOTT.EMP_MV',
     lax  => TRUE);
END;
/

A materialized view can be manually refreshed using the DBMS_MVIEW package.

EXEC DBMS_MVIEW.refresh('EMP_MV');

Rather than using a refresh group, you can schedule DBMS_MVIEW.REFRESH called using the Oracle Scheduler

Cleaning Up

To clean up we must remove all objects.

CONNECT scott/tiger@db2
DROP MATERIALIZED VIEW emp_mv;
DROP DATABASE LINK DB1.WORLD;

BEGIN
  DBMS_REFRESH.destroy(name => 'SCOTT.MINUTE_REFRESH');
END;
/

CONNECT scott/tiger@db1
DROP MATERIALIZED VIEW LOG ON scott.emp;

Aggregations and Transformations

Materialized views can be used to improve the performance of a variety of queries, including those performing aggregations and transformations of the data. This allows the work to be done once and used repeatedly by multiple sessions, reducing the total load on the server.
The following query does an aggregation of the data in the EMP table.

CONN scott/tiger
SET AUTOTRACE TRACE EXPLAIN

SELECT deptno, SUM(sal)
FROM   emp
GROUP BY deptno;

---------------------------------------------------------------------------
| Id  | Operation          | Name | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------
|   0 | SELECT STATEMENT   |      |     3 |    21 |     4  (25)| 00:00:01 |
|   1 |  HASH GROUP BY     |      |     3 |    21 |     4  (25)| 00:00:01 |
|   2 |   TABLE ACCESS FULL| EMP  |    14 |    98 |     3   (0)| 00:00:01 |
---------------------------------------------------------------------------

Create a materialized view to perform the aggregation in advance, making sure you specify the ENABLE QUERY REWRITE clause.

CREATE MATERIALIZED VIEW emp_aggr_mv
BUILD IMMEDIATE 
REFRESH FORCE
ON DEMAND
ENABLE QUERY REWRITE 
AS
SELECT deptno, SUM(sal) AS sal_by_dept
FROM   emp
GROUP BY deptno;

EXEC DBMS_STATS.gather_table_stats(USER, 'EMP_AGGR_MV');

The same query is now rewritten to take advantage of the pre-aggregated data in the materialized view, instead of the session doing the work for itself.

--ALTER SESSION SET QUERY_REWRITE_INTEGRITY = TRUSTED; 
--ALTER SESSION SET QUERY_REWRITE_ENABLED = TRUE;
SET AUTOTRACE TRACE EXPLAIN

SELECT deptno, SUM(sal)
FROM   emp
GROUP BY deptno;

--------------------------------------------------------------------------------------------
| Id  | Operation                    | Name        | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |             |     3 |    21 |     3   (0)| 00:00:01 |
|   1 |  MAT_VIEW REWRITE ACCESS FULL| EMP_AGGR_MV |     3 |    21 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------------------------

Considerations

Before using materialized views and materialized view logs, consider the following:

Populating a materialized view adds load to both servers involved. The source server is queried to capture the data, which is inserted into the destination server. Be sure the additional load does not adversely affect your primary system.
Although materialized view logs improve the performance of materialized view refreshes, they do increase the work needed to perform DDL on the base table. Check the additional work does not adversely affect performance on the primary system.
If regular refreshes are not performed, materialized view logs can grow very large, potentially reducing the performance of their maintenance and blowing tablespace limits.
Depending on the Oracle version and the complexity of the associated query, fast refreshes may not be possible.
When using materialized views to improve performance of transformations and aggregations, the QUERY_REWRITE_INTEGRITY and QUERY_REWRITE_ENABLED parameters must be set or the server will not be able to automatically take advantages of query rewrites. These parameters may be set in the pfile or spfile file if they are needed permanently. Later releases have them enabled by default.

Article taken from http://www.oracle-base.com

Suppose we have a partitioned table and we need gathering data from a specific partition named P_200.
Here is the statement:

SELECT * FROM TABLE_NAME PARTITION (P_200) T
WHERE T.FIELD > 1000

Notes:
Tables can be partitioned into up to 64,000 separate partitions. Any table can be partitioned except those tables containing columns with LONG or LONG RAW data types.

Partitioning Methods

Range Partitioning

Range partitioning maps data to partitions based on ranges of partition key values that you establish for each partition.

When using range partitioning, consider the following rules:

Each partition has a VALUES LESS THAN clause, which specifies a noninclusive upper bound for the partitions. Any binary values of the partition key equal to or higher than this literal are added to the next higher partition.
All partitions, except the first, have an implicit lower bound specified by the VALUES LESS THAN clause on the previous partition.
A MAXVALUE literal can be defined for the highest partition. MAXVALUE represents a virtual infinite value that sorts higher than any other possible value for the partition key, including the null value.

CREATE TABLE sales_range
(salesman_id NUMBER(5),
salesman_name VARCHAR2(30),
sales_amount NUMBER(10),
sales_date DATE)
PARTITION BY RANGE(sales_date)
(
PARTITION sales_jan2000 VALUES LESS THAN(TO_DATE('02/01/2000','DD/MM/YYYY')),
PARTITION sales_feb2000 VALUES LESS THAN(TO_DATE('03/01/2000','DD/MM/YYYY')),
PARTITION sales_mar2000 VALUES LESS THAN(TO_DATE('04/01/2000','DD/MM/YYYY')),
PARTITION sales_apr2000 VALUES LESS THAN(TO_DATE('05/01/2000','DD/MM/YYYY'))
);

List Partitioning

List partitioning enables you to explicitly control how rows map to partitions. You do this by specifying a list of discrete values for the partitioning key in the description for each partition.

CREATE TABLE sales_list
(salesman_id NUMBER(5),
salesman_name VARCHAR2(30),
sales_state VARCHAR2(20),
sales_amount NUMBER(10),
sales_date DATE)
PARTITION BY LIST(sales_state)
(
PARTITION sales_west VALUES('California', 'Hawaii'),
PARTITION sales_east VALUES ('New York', 'Virginia', 'Florida'),
PARTITION sales_central VALUES('Texas', 'Illinois')
PARTITION sales_other VALUES(DEFAULT)
);

Hash Partitioning

Hash partitioning enables easy partitioning of data that does not lend itself to range or list partitioning. It does this with a simple syntax and is easy to implement. It is a better choice than range partitioning when:

You do not know beforehand how much data maps into a given range
The sizes of range partitions would differ quite substantially or would be difficult to balance manually
Range partitioning would cause the data to be undesirably clustered
Performance features such as parallel DML, partition pruning, and partition-wise joins are important

The concepts of splitting, dropping or merging partitions do not apply to hash partitions. Instead, hash partitions can be added and coalesced.

CREATE TABLE sales_hash
(salesman_id  NUMBER(5), 
salesman_name VARCHAR2(30), 
sales_amount  NUMBER(10), 
week_no       NUMBER(2)) 
PARTITION BY HASH(salesman_id) 
PARTITIONS 4 
STORE IN (data1, data2, data3, data4);

The preceding statement creates a table sales_hash, which is hash partitioned on salesman_id field. The tablespace names are data1, data2, data3, and data4.

PL/SQL On the Field

Pages

Thursday, 3 July 2014