Erroneous data in time characteristics that appears in query selections often confuses users. You can resolve this issue by changing a configuration option either in the InfoCube itself or in the InfoObject. However, when reporting from an ODS, this option is not available. Discover how you can resolve this issue by cleaning up the master data of time characteristics and getting rid of those unused records.
In BW, master data tables may have dirty data, or extraneous records, caused by incorrect format of input data. For instance, the date format might be MMYYYY, not YYYYMM. Instead of creating a master data ID for 200510 (October 2005), the system creates an entry for 102005 (May 1020). Another possible cause is incorrect mapping in the transfer or update rules. For example, the system might map the document number to the calendar day because of a programming error. Reporting with dirty data from the source system can contaminate BW.
As you load dirty data into BW, your load may fail. If not, the system assigns a master data ID (SID) to each field value and stores the record in the master data tables. The system never uses these values since they are really errors, but they remain in the SID tables and clutter the system.
These invalid entries in the SID tables generally do not cause problems until users decide to drill down on a characteristic. They then see all the entries in the SID tables (unless they select from possible entries in the InfoCube for the drop-down box, which is not always possible with a line-item dimension or an ODS, as I’ll discuss later). Depending on where the invalid value is on the SID tables, users might have to scroll through many invalid values before they see the value that they want to select.