I hope this is helpful. For example, this is the pattern for a virtual relationship using INTERSECT: The same result can be obtained using TREATAS: The rules of thumb for using these patterns are: If the granularity of the filter propagated is relatively small, you might consider a virtual relationship as a possible alternative to a physical one. I am trying to create a measure TotalExaminationBacklog which counts all the examinationsIDs with the status WAI, VER, APP, HEL and SCH. Specifies an existing relationship to be used in the evaluation of a DAX expression. Consider a model where the Sales table has a day granularity, whereas the table Advertising has a month granularity. If you are used to the INTERSECT pattern, you might find the TREATAS syntax strange, because you must invert the arguments: the first one is the filter context to read, the second one includes the columns. How can I create a slicer in Power BI to ensure I can filter uniquely by: He first started working on Analysis Services in 1998, back when Analysis Services was known as OLAP Services. Asking for help, clarification, or responding to other answers. In this case you have to write an explicit table expression instead of relying on automatic conversion of a logical expression in a table expression made by CALCULATE and CALCULATETABLE when you reference a single column. Specifies cross filtering direction to be used in the evaluation of a DAX expression. In order to obtain such a list, the engine has to execute a table scan. I will edit my post immediately. The most simple form to define a table with just one column is to use {"curly", "braces"}. From hereinafter, we will describe the syntax of the filter arguments in these functions, identified by in the general syntax: CALCULATE (. SUMMARIZE (