Oracle Distributed Systems

Column Groups and Priority Groups

Column groups and priority groups provide interfaces to a variety of built-in conflict resolution techniques and the easiest way to configure your application to resolve conflicts automatically. Let’s look at them one at a time.

Column Groups

A column group is a set of one or more columns associated with a single conflict resolution method. A column cannot belong to more than one column group, and columns that are not explicitly assigned to a column group are members of a shadow column group, which Oracle creates by default and which uses the default conflict resolution methods.

For example, suppose that you want to define a column group for the EMPLOYEES table described in the following table:

Column Name	Type
EMPLOYEE_ID	NUMBER(10)
LAST_NAME	VARCHAR2(30)
FIRST_NAME	VARCHAR2(20)
SOCIAL_SECURITY_NO	VARCHAR2(11)
MARITAL_STATUS	VARCHAR2(1)
HOME_PHONE	VARCHAR2(12)
PERS_GLOBAL_NAME	VARCHAR2(30)
PERS_TIMESTAMP	DATE
MANAGER	NUMBER(3)
DEPT	NUMBER(3)
SAL_GRADE	NUMBER(3)
SENIORITY	VARCHAR2(10)
PAYROLL_GLOBAL_NAME	VARCHAR2(30)
PAYROLL_TIMESTAMP	DATE

Use the procedure DBMS_REPCAT.MAKE_COLUMN_GROUP to define a column group for fields that the personnel site maintains (i.e., LAST_NAME, FIRST_NAME, SOCIAL_SECURITY_NO, MARITAL_STATUS, HOME_PHONE). The application records information about changes to these fields in PERS_GLOBAL_NAME and PERS_TIMESTAMP.

EXECUTE dbms_repcat.make_column_group( - gname => 'HR', - oname => 'EMPLOYEES', - column_group => 'CG_EMP_PERSONNEL', - list_of_column_names => ...