Materialized View

Materialized View
1)creation
1@@@@create a table for a sql statement
@@@original sql statement
sql> set linesize 200
sql> conn sh/sh
sql> ed
SELECT c.cust_id , SUM(amount_sold)
FROM sales s , customers c
WHERE s.cust_id=c.cust_id
GROUP BY c.cust_id
/
SQL> /
7059 rows selected.
Execution Plan
---------------------------------------------------------------------------------------------------------------
| Id  | Operation           | Name          | Rows  | Bytes |TempSpc| Cost (%CPU)| Time     | Pstart| Pstop |
---------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT       |          |  7059 |   103K|       |  2233    (7)| 00:00:27 |       |       |
|   1 |  HASH GROUP BY           |          |  7059 |   103K|    21M|  2233    (7)| 00:00:27 |       |       |
|*  2 |   HASH JOIN           |          |   918K|    13M|       |   459  (10)| 00:00:06 |       |       |
|   3 |    INDEX FAST FULL SCAN| CUSTOMERS_PK | 55500 |   270K|       |    28    (4)| 00:00:01 |       |       |
|   4 |    PARTITION RANGE ALL |          |   918K|  8973K|       |   419    (8)| 00:00:06 |     1 |    28 |
|   5 |     TABLE ACCESS FULL  | SALES          |   918K|  8973K|       |   419    (8)| 00:00:06 |     1 |    28 |
---------------------------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
   2 - access("S"."CUST_ID"="C"."CUST_ID")

Statistics
----------------------------------------------------------
second:(with cache)
       1840  consistent gets
      0  physical reads

@@@
sql> ed
CREATE TABLE cust_sales_sum AS
SELECT c.cust_id , SUM(amount_sold) AS amount
FROM sales s , customers c
WHERE s.cust_id=c.cust_id
GROUP BY c.cust_id
/
SQL> select * from cust_sales_sum;
7059 rows selected.
Execution Plan
------------------------------------------------------------------------------------
| Id  | Operation      | Name       | Rows  | Bytes | Cost (%CPU)| Time       |
------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |           |  7059 |   179K|     6   (0)| 00:00:01 |
|   1 |  TABLE ACCESS FULL| CUST_SALES_SUM |  7059 |   179K|     6   (0)| 00:00:01 |
------------------------------------------------------------------------------------
Note
-----
   - dynamic sampling used for this statement

Statistics
----------------------------------------------------------
     48  recursive calls
    510  consistent gets
     15  physical reads



2@@@@create a materialized view for that sql statement
sql> conn sh/sh
sql> ed
CREATE MATERIALIZED VIEW cust_sales_mv
ENABLE QUERY REWRITE AS
 SELECT c.cust_id , SUM(amount_sold) AS amount
 FROM  sales s, customers c
 WHERE s.cust_id=c.cust_id
 GROUP BY c.cust_id
/
SQL> /
Materialized view created.


sql> ed
 SELECT c.cust_id , SUM(amount_sold) AS amount
 FROM  sales s, customers c
 WHERE s.cust_id=c.cust_id
 GROUP BY c.cust_id
/
SQL> /
7059 rows selected.
Execution Plan
----------------------------------------------------------------------------------------------
| Id  | Operation             | Name         | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT         |             |    7059 |     179K|       6   (0)| 00:00:01 |
|   1 |  MAT_VIEW REWRITE ACCESS FULL| CUST_SALES_MV |    7059 |     179K|       6   (0)| 00:00:01 |
----------------------------------------------------------------------------------------------
Note
-----
   - dynamic sampling used for this statement

Statistics
----------------------------------------------------------
second:(with cache)
      5  recursive calls
    506  consistent gets
      0  physical reads

first:(without cache)
       4082  recursive calls
      0  db block gets
       1679  consistent gets
      0  physical reads

@@@summary:
@@@materialized view is convenient for refresh and rewrite for sql.
@@@create table is simlar to create materialized view.


2)maintainance
1@@@@refresh
@@@use EM => adminstration => materialized view
<1> on demand (manual)
<2> on commit
<3> automatically (schedual)
@@@evaluation: auto is most useful.

2@@@@use "dbms_mview" package
@@@refresh a specific materialized views.
SQL> exec dbms_mview.refresh('cust_sales_mv');

@@@refresh all materialized view which depend on a set of base table.
SQL> VARIABLE fail NUMBER;
SQL> exec dbms_mview.refresh_dependent(:fail,'customers,sales');
PL/SQL procedure successfully completed.
SQL> print fail

@@@
@@@on demand
@@@use EM => adminstration => materialized view , it could be changed easily.
SQL> select refresh_mode, staleness, update_log
     from dba_mviews
     where mview_name='CUST_SALES_MV' and owner='SH';
REFRESH_MODE       STALENESS              UPDATE_LOG
------------------ ------------------------------ ------------------------------
DEMAND           FRESH


@@@
@@@on commit
@@@use EM => adminstration => Materialized view => 2M
ALTER MATERIALIZED VIEW SH.CUST_SALES_MV REFRESH ON COMMIT

@@@
@@@on schedual
ALTER MATERIALIZED VIEW SH.CUST_SALES_MV NOLOGGING REFRESH START WITH
to_date('07-03-2012 15:17:36','MM-dd-yyyy hh24:mi:ss')
NEXT sysdate + 3/1440


3@@@@query rewrite
<1>hint "REWRITE_OR_ERROR"
@@@
@@@check the whether the sql statement rewrited or not.
@@@if it is not rewrited, echo error to user,
@@@if it is rewrited, echo nothing
sql> ed
CREATE MATERIALIZED VIEW cust_orders_mv
ENABLE QUERY REWRITE AS
SELECT c.customer_id, SUM(order_total) AS amt
FROM  oe.orders s, oe.customers c
WHERE s.customer_id = c.customer_id
GROUP BY c.customer_id
/
SQL> /
Materialized view created.

sql> ed
SELECT /*+ REWRITE_OR_ERROR */
       c.customer_id,
       SUM(order_total) AS amt
FROM oe.orders s, oe.customers c
WHERE s.customer_id = c.customer_id
GROUP BY c.customer_id
/
sql> /
FROM oe.orders s, oe.customers c
        *
ERROR at line 4:
ORA-30393: a query block in the statement did not rewrite
      

<2>query rewrite example
@@@
@@@create a materialized view cost_per_year_mv
SQL> show parameter query
NAME                     TYPE                   VALUE
------------------------------------ --------------------------------- ---------
query_rewrite_enabled             string                   true
query_rewrite_integrity          string                   enforced

SQL> show parameter mode      @@@<==the parameter must be CBO
NAME                     TYPE                   VALUE
------------------------------------ --------------------------------- ----------
optimizer_mode                 string                   ALL_ROWS

sql> ed
CREATE MATERIALIZED VIEW cost_per_year_mv
ENABLE QUERY REWRITE AS
SELECT t.week_ending_day,
       t.calendar_year,
       p.prod_subcategory,
       sum(c.unit_cost) AS dollars
FROM costs c, times t, products p
WHERE c.time_id = t.time_id
AND   c.prod_id = p.prod_id
GROUP BY t.week_ending_day,
         t.calendar_year,
         p.prod_subcategory
/

sql> ed             @@@<==here query_rewrite enabled and work,"MAT_VIEW REWRITE ACCESS FULL"
SELECT t.week_ending_day,
       t.calendar_year,
       p.prod_subcategory,
       sum(c.unit_cost) AS dollars
FROM costs c, times t, products p
WHERE c.time_id = t.time_id
AND   c.prod_id = p.prod_id
GROUP BY t.week_ending_day, t.calendar_year,
         p.prod_subcategory
HAVING sum(c.unit_cost) > 10000   @@@<==just a "having statement" more than the mv query.
/
174 rows selected.
Execution Plan
-------------------------------------------------------------------------------------------------
| Id  | Operation             | Name        | Rows    | Bytes | Cost (%CPU)| Time    |
-------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT         |            |   174 | 10788 |     7   (0)| 00:00:01 |
|*  1 |  MAT_VIEW REWRITE ACCESS FULL| COST_PER_YEAR_MV |   174 | 10788 |     7   (0)| 00:00:01 |
-------------------------------------------------------------------------------------------------
      0  recursive calls
      0  db block gets
     35  consistent gets

sql> ed          @@@<==here "MAT_VIEW ACCESS FULL"
SELECT week_ending_day,
       prod_subcategory,
       dollars
FROM cost_per_year_mv
WHERE calendar_year = '1999'
AND  dollars > 10000
/
9 rows selected.
Execution Plan
-----------------------------------------------------------------------------------------
| Id  | Operation         | Name        | Rows    | Bytes | Cost (%CPU)| Time    |
-----------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT     |            |     9 |   558 |     7   (0)| 00:00:01 |
|*  1 |  MAT_VIEW ACCESS FULL| COST_PER_YEAR_MV |     9 |   558 |     7   (0)| 00:00:01 |
-----------------------------------------------------------------------------------------
      0  recursive calls
      0  db block gets
     24  consistent gets

4@@@@fast refresh which base on materialized view log
@@@if the base table often was modified by DML, may be you could create
@@@the mview log on it.
@@@use EM => adminstration => materialized view log
@@@detail referencing
http://docs.oracle.com/cd/B12037_01/server.101/b10759/statements_6003.htm

@@@
SQL> ed
CREATE MATERIALIZED VIEW LOG ON oe.product_information
   WITH ROWID, SEQUENCE (list_price, min_price, category_id)
   INCLUDING NEW VALUES;
SQL> /
Materialized view log created.

SQL> ed
CREATE MATERIALIZED VIEW oe.products_mv
   REFRESH FAST ON COMMIT
   AS SELECT SUM(list_price - min_price), category_id
         FROM product_information
         GROUP BY category_id;
SQL> /
Materialized view created.