A data source is any tabular structure. You get to see data source values after you load a file or send a query to a service that returns results (for example, another Oracle Business Intelligence system or a database).
A data source can contain any of the following:
Match columns: These contain values that are found in the match column of another source, which relates this source to the other (for example, Customer ID or Product ID).
Attribute columns: These contain text, dates, or numbers that are required individually and aren’t aggregated (for example, Year, Category Country, Type, or Name).
Measure columns: These contain values that should be aggregated (for example, Revenue or Miles driven).
You can analyze a data source on its own, or you can analyze two or more data sources together, depending on what the data source contains.
When you save a project, the permissions are synchronized between the project and the external sources that it uses. If you share the project with other users, then the external sources are also shared with those same users.
Working with Sources with no Measures
Note the following if you’re working with sources with no measures.
If a table has no measures, it’s treated as a dimension. Note the following criteria for extending a dimension:
Matches can be between one or composite columns. An example of a one column match is that product key matches product key. For composite columns, an example is that company matches company and business unit matches business unit.
All other columns must be attributes.
Dimension tables can be matched to other dimensions or they can be matched to tables with measures. For example, a table with Customer attributes can be matched to a table with demographic attributes provided both dimensions have unique Customer key columns and Demographic key columns.
Working with Sources with Measures
Note the following if you are working with sources with measures.
You can match tables with measures to other tables with a measure, a dimension, or both.
When you match tables to other tables with measures, they don’t need to be at the same grain. For example, a table of daily sales can be matched to a table with sales by Quarter if the table with the daily sales also includes a Quarter column.
Working with Matching
If you use multiple sources together, then at least one match column must exist in each source. The requirements for matching are:
The sources contain common values (for example, Customer ID or Product ID).
The match must be of the same data type (for example, number with number, date with date, or text with text).