Add Float64 to ColumnType

[brunto]

This is a solution for the issue: #636

[matthewmcgarvey]

@brunto we're not going to remember why this type union was added. Could you add a spec that would fail before the change and this fixes?

[jwoertink]

Nice! Thanks for adding these specs. It's good to have them in.

[brunto]

It's not finished I have one more test to add.

[brunto]

Ok now I'm done.

[jwoertink]

@brunto now that you added that DOUBLE, I'm not sure if that type actually gets checked. I think it's just at runtime when you query on it, so you'd have to now add a spec that checks you can query on a DOUBLE column.

[jwoertink]

I think the rest of it is fine. I'm not sure the best place to put the spec, but you can probably just add it here and say that we can query float columns defined as DOUBLE.

[brunto]

@jwoertink I just saw the "bucket" table. Maybe it's better to put the "Double" columns inside Bucket instead of Business?

[jwoertink]

@brunto the buckets was used for Array stuff. I guess if you wanted to test a postgres array of double, then you can add an additional one there too, but otherwise what you have in business is fine.

[jwoertink]

Ok, I pulled this in to one of my projects, and it seems to work. @brunto if you get a sec, can you rebase to make sure everything is good?

[brunto]

I rebased the code but I have one more spec to push.

[brunto]

@jwoertink Ok all is good now.

[jwoertink]

woo hoo! Thanks for tackling this. 🎉