The deal with two nodes with the same name you could refer to them by their position. In this code:

SELECTx.value('(/CommonEventData/Row[1]/CallSource)[1]','NVARCHAR(50)') AS CallSource,

x.value('(/CommonEventData/Row[1]/CallerCityStateZipCode)[1]','NVARCHAR(50)') AS CallerCityStateZipCode,

x.value('(/CommonEventData/Row[1]/CallerName)[1]','NVARCHAR(50)') AS CallerName,

x.value('(/CommonEventData/Row[1]/CallerPhone)[1]','NVARCHAR(50)') AS CallerPhone,

x.value('(/CommonEventData/Row[2]/Beat)[1]','NVARCHAR(50)') AS Beat,

x.value('(/CommonEventData/Row[2]/MultiEventId)[1]','NVARCHAR(50)') AS MultiEventId,

x.value('(/CommonEventData/Row[2]/PrimaryUnitId)[1]','NVARCHAR(50)') AS PrimaryUnitId

FROM (SELECT @TXML) x(x)

In this code, Row[1] is the First instance of the "row" node. Row[2] is the Second instance of the 'row' node.

Since the children of row are unique (e.g. callerName, CallerPhone, etc only appear once) you can refer only to row and let the XML parser sort it out like so...

SELECTx.value('(/CommonEventData/Row/CallSource)[1]','NVARCHAR(50)') AS CallSource,

x.value('(/CommonEventData/Row/CallerCityStateZipCode)[1]','NVARCHAR(50)') AS CallerCityStateZipCode,

x.value('(/CommonEventData/Row/CallerName)[1]','NVARCHAR(50)') AS CallerName,

x.value('(/CommonEventData/Row/CallerPhone)[1]','NVARCHAR(50)') AS CallerPhone,

x.value('(/CommonEventData/Row/Beat)[1]','NVARCHAR(50)') AS Beat,

x.value('(/CommonEventData/Row/MultiEventId)[1]','NVARCHAR(50)') AS MultiEventId,

x.value('(/CommonEventData/Row/PrimaryUnitId)[1]','NVARCHAR(50)') AS PrimaryUnitId

FROM (SELECT @TXML) x(x)

What I posted above is the most simplified.