![sap hana studio null handling sap hana studio null handling](https://blogs.sap.com/wp-content/uploads/2015/12/pastedimage_0_857483.png)
Shouldn’t this join type never drop data in the fact? The answer is no, even in a left outer join, when you apply a filter on the right table an inner join is effectively executed. When the user now implements a filter on “attribute 1”, the null value is dropped and the $100 “disappears”, which can cause some heartburn for the average user, “where did my sales go?!” The model is built with a fact and dimension, having a left outer join with a 1:n relationship. Real World Problem: User observes that while querying a given model in HANA, there is $300 in total sales by “attribute 1”, with $100 of that falling into a null (“?”) “attribute 1”.
![sap hana studio null handling sap hana studio null handling](https://miro.medium.com/max/712/1*cgNGQYxiBqHa0Qm2swRy8A.jpeg)
If I had to guess, this is to ensure this behavior does not occur! Interesting to note, in SAP BW, master data tables/InfoObjects always have an entry with a blank row. This simply describes an easy way to solve using HANA based components. Just as a side note – this issue is not specific to HANA and can (and does) need to be addressed in whatever database your solution may be implemented in. In general, we want to avoid returning null attribute values for a given entry in a fact. Purpose: Describe a method to address a common DW/BI problem of not having a matching row in a dimension for a given fact where the fact column is blank (” ) whitespace.