In source file preview screen you can enter data cube name where you would like to import your data. If this is new data set then enter new cube name (e.g. Sales cube if you are importing sales transactions). You can also import new file in existing data cube - in this case you can use either existing cube dimensions and measures or specify new ones (which will be added to existing cube during import).
It is recommended that CSV file has header row with column names - eazyBI will try to detect if your CSV file has header row or not. If eazyBI has detected it incorrectly then please change number of header rows (e.g. specify 0 if you do not have any header row).
eazyBI will try to detect each source file column data type based on first 100 rows of source file. Detected column types (string, integer,decimal, date or datetime) are shown below each column name. If some column type is detected incorrectly then you can change it (e.g. change from integer to string if this column might contain also non-integer values).
Column mapping to dimension and dimension level
Next you need to map source file columns to cube dimensions and measures.
By default eazyBI will suggest to map string columns to dimension with name which is derived from column name (underscores replaced with spaces, each word is capitalized). If necessary then you can enter different dimension name that you would like to use in eazyBI.
You can map each column to separate dimension but sometimes it would be better to map several columns to the same dimension but different hierarchy levels. E.g. in this example it would be better to map country, state_province, city and fullname to the same Customers dimension with hierarchy levels Country (highest hierarchy level), State Province (one country expands to several states or provinces), City (one state or province expands to several cities), Customer Name (one city expands to several individual customers). It is important that in source file hierarchy level columns are always in the sequence from highest hierarchy level to lowest hierarchy level.
When you map several columns to the same dimension then you can select existing dimension name from list for next columns (delete suggested default dimension name to see list of available dimension names).
If you map several columns to the same dimension then it is mandatory to specify dimension hierarchy level for each column. By default hierarchy level name will be generated from column name (in the same way as default dimension name). If necessary then you can enter different dimension hierarchy level name.
Ignore columns
If you would like to ignore some columns during import (do not want that these columns are imported) then you should leave dimension selection blank for this column.
Time dimension
Date or datetime column by default will be mapped to Time dimension. Time dimension after import will automatically create year, quarter,month and day hierarchy levels as well as additional weekly hierarchy with year, week and day hierarchy levels. Therefore you should not provide any dimension hierarchy level information in source file mapping as time hierarchies will be created automatically.
You should use one of the following date formats in your CSV file that eazyBI will recognize as date format and you will be able to import it as Time dimension:
- YYYY-MM-DD
- YYYY.MM.DD or DD.MM.YYYY
- YYYY/MM/DD or MM/DD/YYYY
The following datetime formats are recognized (currently only date portion of datetime value will be imported in Time dimension):
Measures and integer and decimal columns
Numeric columns (integer and decimal data types) by default will be mapped to Measures and you need to specify measure names for these columns (by default derived from column names). Measures are values that you would like to analyze across other dimensions and which will be aggregated by other dimensions hierarchy levels.
If you have other integer values (most typically e.g. identifier columns like customer ID or order ID etc.) in source file columns then do not map them to Measures - either enter corresponding dimension name or leave dimension name blank to ignore them during import.
