Sometimes the facts from an operational business process arrive minutes, hours, days, or weeks before the associated dimension context. For example, in a real-time data delivery situation, an inventory depletion row may arrive showing the natural key of a customer committing to purchase a particular product. In a real-time ETL system, this row must be posted to the BI layer, even if the identity of the customer or product cannot be immediately determined.  In these cases, special dimension rows are created with the unresolved natural keys as attributes.  Of course, these dimension rows must contain generic unknown values for most of the descriptive columns; presumably the proper dimensional context will follow from the source at a later time. When this dimensional context is eventually supplied, the placeholder dimension rows are updated with type 1 overwrites. Late arriving dimension data also occurs when retroactive changes are made to type 2 dimension attributes. In this case, a new row needs to be inserted in the dimension table, and then the associated fact rows must be restated.