Whenever possible, the records have the same structure: each record contains the star identification, the source and the data. White space are used as separators and no padding is used for missing data. The fields of each datatype are listed in Appendix E. The star identification is the main key to access the data, but it is also possible to use filters based on the data sources or astrophysical parameters.
Most data files are 1 to n relations because several sources of data may exist for each star. Files with mean or selected data are 1 to 1 relations. The first aim of the database and of the data analysis facilities developed is to pass from the 1 to n state to the 1 to 1 state. This means that at the end of the analysis process, only one value of each parameter will remain for each star.