Decimal accuracy in report view versus data view

Problem with very small amounts being added to my decimal values (i.e. .000000000000050). This is happening with my SQL data but I can repro it without an external datasource. Hopefully someone will know a secret I'm missing. Here goes:

• Open a blank Power BI report, click Enter Data from the Home tab. Under Column 1, Row 1; enter A and hit Tab. Column 2 appears; enter 754.25 and hit Tab. Column 3 appears; enter 739.16 and hit Tab. Column 4 appears; enter 15.09. On the next several rows, continue entering the following:
  • Row 2; B, 975.32, 955.81, 19.51
  • Row 3; C, 192787.73, 177609.54, 0
  • Row 4; D, 905.17, 905.17, 0

• Click to open Data view. You have a Table with 4 rows.
• Click Report view, add a Table visualization, and add Column1, Column2, Column3, and Column4. Right click the Table name and select New column. Define it as: Column = [Column2] - [Column3] - [Column4]. You will see three of the rows show Column as zero. Add Filter on Column as is not 0 and hit Apply. Hmmmm...only one row disappeared. Why?
• Click the Column field, click the Modeling tab, and change the Formatting Decimal Places to 15 so you can see what's what. Aha! Now you see the deception: Row 1 (A) is 0.000000000000032 and row 2 (B) is 0.000000000000103. Also notice row 3 (C) is 15178.190000000002000. Set the Decimal Places for Column2, Column3, and Column4 to 15. You'll see many of them with the same problem.
• Click Data view and notice all the data appears correct here, except for Column.

You can guess how this is wreaking havoc in a complex report that's based on whether or not Column is zero or not.

I can change the definition for Column to: Column = Round([Column2] - [Column3] - [Column4], 6) and that resolves it for this filter, but there are many other tests that are not as straight forward.

This seams similar to this post: Power Query decimal precision problem - does not get to 0 when negative values equal positive values , but the solution there seems to be a band-aid to a bigger problem. Or maybe I don't know how to efficiently apply that solution to my situation.

I'd like to know if there's a quicker, more fundamental solution, before going through all of my calculations to add ROUND() all over, while not being entirely sure it will be effective in every case.