Use the Explore option in the DSV ...
Try right-clicking the table and selecting Explore Data ... this will only return a sample of the rows for a large table, but the viewer offers several tabs that might assist you in locating patterns - which, in turn, might lead to why rows are being omitted.
While you're in the DSV, check those joins again, as well ... and look closely at data types for the joined columns ... you say that the data pulls fine in a query, so something is happening on the induction side of the cube that is limiting intake to a subset ...
Let us know how it goes ...
Bill
Not Using Surrogate Keys?
Am I understanding that you're joining on timestamps and not surrogate keys? The latter is a better approach. A quick and easy table can also be generated through the "push down" approach I demo in my article at:
http://www.databasejournal.com/featu...le.php/3658391
--- even if you run through the wizard and create a simple cube / use a sample cube just to generate the time dim table (and scrap the rest) - I do this all the time to generate the myriad rows (it really gets myriad when you're using time units less than single days ...) quickly.
Yes, I can see how the time stamp joins might cause issues - if I'm understanding you (forgive me if not ...). Let us know if you can approach it another way, such as the aforementioned.
Good Luck.
Bill
What are You Using as Surrogates for Time Dim?
What are your surrogate keys - integers?
Thanks.
Bill
Differing Data Types on the Surrogate Keys?
Not saying that it's the core problem, by any means, but why would you have surrogate keys with differing data types across tables of the star?
Bill