Add turbodbc

Source: https://github.com/blue-yonder/turbodbc

I would volunteer to package it but have some open questions:

• It would depend on unixodbc which is not yet packaged in conda-forge but externally: Should I also add a package for it?
• It may also be helpful to provide two versions of turbodbc, one built with iodbc and one with unixodbc. This would be very roughly similiar to selecting a specific numpy version but I have no idea how this could be done or where I would have to search for documentation. Any pointers are welcome here. If this is not possible, I would default to only the unixodbc based version.

[wesm]

Yes, you should add a unixodbc package.

For the iodbc package, it’s tricky but you could implement this using conda “features” – e.g. NumPy navigates this with the mkl, non-MKL, and openblas variants (e.g. https://github.com/conda-forge/numpy-feedstock/blob/master/recipe/meta.yaml#L19). This is more sparsely documented than you would like

[jakirkham]

Also please raise this sort of issue at the feedstock in the future. Thanks.