Oracle9i OLAP Services Concepts and Administration Guide for Windows Release 1 (9.0.1) for Windows Part Number A90371-01 |
|
Configuring Your Data Warehouse, 4 of 11
The following list summarizes the database requirements:
OLAP metadata considers Time dimensions as distinct from other dimensions. When you specify a dimension in OLAP management, you must identify whether it is a Time dimension. A Time dimension has special attributes that support both regular and irregular time periods.
Regular time periods, such as weeks, months, and years, are evident on standard calendars. Typically, they neither overlap nor have gaps between them. Irregular time periods, such as promotional schedules and seasonal time periods, are not evident on standard calendars. They often overlap (even to the extent that one time period is a subset of another time period) or have gaps between them.
The Time dimension table must contain the following columns:
The following table describes an example dimension table in a star schema.
The following tables describe example dimension tables in a snowflake schema. The first table defines weeks, which is the lowest level of time data.
A second table defines quarters.
A third table defines years.
The OLAP API supports native Java data types. It does not support the following Oracle data types: BLOB, CLOB, NCLOB, RAW, and LONG RAW. Do not create measures from facts with these unsupported data types.
Note: The OLAP DML supports CLOB and NCLOB data types. If your application requires this type of data, then you should investigate a MOLAP or HOLAP solution. Search for "SQL (FETCH)" in OLAP DML Help for additional information about supported data types.
|
Copyright © 1996-2001, Oracle Corporation. All Rights Reserved. |
|