Purpose
This statement is used to create a materialized view log.
A materialized view log (mlog) records incremental updates to the base table to support the fast refresh feature of materialized views. An mlog is a record table that tracks changes in the base table and applies these changes to the corresponding materialized view to enable fast refresh.
Note
In OceanBase Database, the mlog does not support specifying partitions. The partitions of the mlog are bound to those of the base table.
Privilege requirements
To create a materialized view log, you must have the CREATE TABLE and SELECT privileges on the base table. For more information about the privileges in OceanBase Database, see Privilege types in MySQL mode.
Syntax
CREATE [OR REPLACE] MATERIALIZED VIEW LOG ON [database.] table
[parallel_clause]
[with_clause]
[mv_log_purge_clause];
parallel_clause:
NOPARALLEL
| PARALLEL integer
with_clause:
WITH [ {PRIMARY KEY | ROWID | SEQUENCE}
[ { , PRIMARY KEY | , ROWID | , SEQUENCE }]... ]
(column_name [, column_name]...)
[new_values_clause]
new_values_clause:
{INCLUDING | EXCLUDING} NEW VALUES
mv_log_purge_clause:
PURGE {IMMEDIATE [ SYNCHRONOUS ]
| START WITH datetime_expr [NEXT datetime_expr]
| [START WITH datetime_expr] NEXT datetime_expr
}
Parameters
| Parameter | Description |
|---|---|
| OR REPLACE | Optional. Specifies OR REPLACE to indicate that if the corresponding mlog already exists, a new mlog will be created based on the definition and replace the existing mlog. During the recreation of the mlog, the normal read and write operations of the materialized view and base table are not affected. |
| database. | Optional. Specifies the database where the base table of the materialized view log is located. If you omit database., the base table is assumed to be in your own database. |
| table | Specifies the name of the base table corresponding to the materialized view log. |
| parallel_clause | Optional. Specifies the degree of parallelism (dop) for creating the materialized view table, that is, the level of parallel processing for the materialized view log. For more information, see parallel_clause below. |
| with_clause | Optional. Specifies the auxiliary columns included in the materialized view log. It indicates whether to record the primary key (PRIMARY KEY) and row identifier (ROWID) when the base table changes. You can also use this clause to add a sequence (SEQUENCE) to provide additional sorting information for the materialized view log. For more information, see with_clause below. |
| column_name | Optional. Specifies the name of the column whose value is recorded in the materialized view log for all modified rows. |
| new_values_clause | Optional. Specifies whether to record both the old and new values in the materialized view log for update operations. For more information, see new_values_clause below. |
| mv_log_purge_clause | Optional. Specifies the time when data in the materialized view log is purged. For more information, see mv_log_purge_clause below. |
parallel_clause
NOPARALLEL: the default configuration. The parallelism is1.PARALLEL integer: specifies the parallelism. The value ofintegermust be greater than or equal to1.
with_clause
PRIMARY KEY: indicates that the materialized view log records the primary key columns of the base table. This option is effective only for tables with a primary key. If you do not specify this option, the system automatically adds this attribute to the materialized view log, meaning that the materialized view log will default to including the primary key columns.ROWID: indicates that the materialized view log records the row identifiers (ROWID) of the base table. If you do not specify this option, the system automatically adds this attribute to the materialized view log, meaning that the materialized view log will default to including theM_ROW$$column.SEQUENCE: indicates that the materialized view log records the sequence numbers (seq_no) of multiple-row updates within a transaction. The system automatically adds this attribute to the materialized view log, meaning that the materialized view log will default to including theSEQUENCE$$column.
new_values_clause
INCLUDING: the default setting. Indicates that both the old and new values can be recorded in the materialized view log. If you want to enable fast refresh for the materialized view, you must specifyINCLUDING NEW VALUES.EXCLUDING: indicates that new values are not recorded in the materialized view log. Do not useEXCLUDING NEW VALUES, or an error will be returned.
mv_log_purge_clause
Notice
The corresponding data in the materialized view log will be purged only after the materialized view is successfully refreshed.
IMMEDIATE: specifies that the corresponding materialized view log is purged immediately after the materialized view is refreshed. The default value isSYNCHRONOUS.SYNCHRONOUS: optional. Specifies that the purge is executed synchronously.
START WITH datetime_expr [NEXT datetime_expr]:START WITH datetime_expr: specifies the initial purge time for the materialized view log.[NEXT datetime_expr]: optional. Specifies the next purge time for the materialized view log. It is used to set the next purge time.
[START WITH datetime_expr] NEXT datetime_expr: if you do not specify theSTART WITH datetime_exprparameter but specify theNEXT datetime_exprparameter, the initial purge time for the materialized view log is set to the value of theNEXT datetime_exprparameter.
Notice
Make sure that the START WITH datetime_expr and NEXT datetime_expr expressions are set to future dates and times, otherwise an error will be returned.
We recommend that you use sysdate() to represent the current time in the current time zone. Here is an example of a time expression:
START WITH sysdate() NEXT sysdate() + INTERVAL 1 DAY
The preceding clause indicates that expired materialized view log records are purged every 1 day starting from the current time (sysdate()).
Examples
Create a table named
test_tbl1.obclient> CREATE TABLE test_tbl1 (col1 INT PRIMARY KEY, col2 VARCHAR(20), col3 INT);Create a materialized view log on the
test_tbl1table. Set the parallelism for processing the materialized view log to5, record changes to thecol2column, and include both the old and new values of the changed column. Configure the materialized view log to start from the current date and clear expired records every1day.obclient> CREATE MATERIALIZED VIEW LOG ON test_tbl1 PARALLEL 5 WITH SEQUENCE(col2) INCLUDING NEW VALUES PURGE START WITH sysdate() NEXT sysdate() + INTERVAL 1 DAY;View the information of the materialized view log on the
test_tbl1table.obclient> DESC mlog$_test_tbl1;The returned result is as follows:
+------------+-------------+------+------+---------+-------+ | Field | Type | Null | Key | Default | Extra | +------------+-------------+------+------+---------+-------+ | col1 | int(11) | NO | PRI | NULL | | | col2 | varchar(20) | YES | | NULL | | | SEQUENCE$$ | bigint(20) | NO | PRI | NULL | | | DMLTYPE$$ | varchar(1) | YES | | NULL | | | OLD_NEW$$ | varchar(1) | YES | | NULL | | +------------+-------------+------+------+---------+-------+ 5 rows in setChange the materialized view log on the
test_tbl1table to record changes to thecol2andcol3columns.obclient> CREATE OR REPLACE MATERIALIZED VIEW LOG ON test_tbl1 PARALLEL 5 WITH SEQUENCE(col2 ,col3) INCLUDING NEW VALUES PURGE START WITH sysdate() NEXT sysdate() + INTERVAL 1 DAY;View the information of the materialized view log on the
test_tbl1table.DESC mlog$_test_tbl1;The returned result is as follows:
+------------+-------------+------+------+---------+-------+ | Field | Type | Null | Key | Default | Extra | +------------+-------------+------+------+---------+-------+ | col1 | int(11) | NO | PRI | NULL | | | col2 | varchar(20) | YES | | NULL | | | col3 | int(11) | YES | | NULL | | | SEQUENCE$$ | bigint(20) | NO | PRI | NULL | | | DMLTYPE$$ | varchar(1) | YES | | NULL | | | OLD_NEW$$ | varchar(1) | YES | | NULL | | +------------+-------------+------+------+---------+-------+ 6 rows in set
