Question

I have two fields in my database that are defined as floats

[Longitude] [float] NULL
[Latitude] [float] NULL

I have added them to my linq2sql dbml and they show as floats as follows

I can see in my database there is the value -3.6607431 for Longitude however when I profile the sql (using sql express and a custom profiler tool) I see that the actual value passed in queries is 3.6607429981231689.

It seems to have lost the negative sign and had extra values added.

Is there something I’m missing does a float in SQL not map to a float like this in the dbml or is there another issue going on

Answer 1

A value of -3.6607431f in C# is actually exactly -3.6607429981231689453125 – i.e. that’s the exactly-representable float value which is closest to -3.6607431. So I think that answers the precision part – it’s just a natural part of binary floating point. Basically, you shouldn’t worry about that – unless you need precise decimal values, in which case you should map to decimal instead of float (and update your database schema appropriately).

The sign issue is somewhat different – and my guess is that you’ve misdiagnosed it somewhere, or you’ve got some code which negates things. I very much doubt that that’s a database problem.

It’s not clear from your description whether you’re fetching results or passing a parameter in a query (you talk about a “value passed in queries” which could mean either way). If you could show the code involved and how you’re diagnosing the problem, that would help.