Allow filtering of data on load from file-based sources like excel/csv

When reading data from file or other non-relational type sources, support the ability to filter the data as part of the load process.

I see this addressing 2 different needs.  The first is simply being able to support only loading a subset of the data for simple reasons like only wanting the current year data passed through like we can do with any relational type database.  The second part is for cases where there may be multiple record types or non-data rows such as invalid rows, multiple header row occurrences or formatted reports with total lines and such.  The filtering should support simple things like column1='abcd' and more complex options like ignoring an entire row if a certain text string occurs like "Total for abcd".

You can obviously filter the data after its been pulled into Spotfire but that creates an issue where the filter has to be explicitly set in the DXP which means it can also be unset.  Another example would be having a single excel sheet that contains actuals and budget data where I want to import it as 2 separate data tables (actual and budget).  There is no easy way of doing this today in Spotfire.  Another option would be creating something like "hard" filter that is enabled outside the traditional filtering (a setting on the Data Table Properties possibly like with On Demand) which can't easily be overriden.