You use the Header tab to specify the location of data-source header records.
In the “Number of lines to skip” box, you specify how many header lines (counting from the top) can be skipped when data is loaded. You can skip from 0 to 999,999 records. You should skip all header records.
In the boxes of the “Data source header records” group, you identify header locations:
Record containing header names—Enter the number of the record that specifies the dimensions and members that are used as data-source headers. The specified dimensions and members are used for each data-source record. Only one member from each dimension can be specified. For example, if the sample header name contains “Jan, Sales,” each record in the data source is mapped to the Jan member of the Year dimension and the Sales member of the Measures dimension.
Record containing data load field names—Enter the number of the record that contains the data load field names of the data source. Data load field names are dimension or member names that tell Essbase how the data load fields of the data source are ordered. A sample data load field header is Year, Measures, Market, Product, Scenario.
Record containing dimension building field names—Enter the number of the record that contains the dimension-build field names of the data source. Dimension-build, field-name headers consist of dimension names and field types, such as Level 0, Product, Level 1, Product, Level 2, Product.
“Defining Header Records” in the Oracle Essbase Database Administrator's Guide
“Data Source Headers” in the Oracle Essbase Database Administrator's Guide
“Valid Data Source Header Field Types” in the Oracle Essbase Database Administrator's Guide