I've been using my own implementation of a Decimal type -- maybe it should be included as a standard type? from graphene.types import Scalar from graphql.language import ast class Decimal(Scalar): ...

Background - float type can't store all decimal numbers exactly For numbers with a decimal separator, by default Python uses float and Pandas uses numpy float64. Internally float types use a base 2 ...

So C# has decimals, numeric real literals which are stored as doubles unless you suffix it with the type (M or m for decimal)... I don't quite understand the significance of this. Also of course ...