[Solved] Why does Float has a constructor that allows double as an argument while Double doesn’t have a constructor with float as an argument in JAVA?

Question

I think you’ve said why in your question, but just for clarity:

Double(double) works just fine if you pass it a float, so there’s no need for Double(float). This is because in a constructor invocation (JLS§5.3), a widening primitive conversion (JLS§5.1.2) is allowed. float to double is a widening primitive conversion.

But Float(float) would not be just fine if you passed it a double; you’d get a warning of a loss of precision because that’s a narrowing primitive conversion (JLS§5.1.3). So by providing a Float(double) constructor which explicitly states it will lose precision, the API makes it possible to write code calling Float‘s constructor with a double without the warning. (You could argue whether that’s a good thing vs. having callers do the explicit (float) cast when making the call, but…)

Accepted Answer

I think you’ve said why in your question, but just for clarity:

Double(double) works just fine if you pass it a float, so there’s no need for Double(float). This is because in a constructor invocation (JLS§5.3), a widening primitive conversion (JLS§5.1.2) is allowed. float to double is a widening primitive conversion.

But Float(float) would not be just fine if you passed it a double; you’d get a warning of a loss of precision because that’s a narrowing primitive conversion (JLS§5.1.3). So by providing a Float(double) constructor which explicitly states it will lose precision, the API makes it possible to write code calling Float‘s constructor with a double without the warning. (You could argue whether that’s a good thing vs. having callers do the explicit (float) cast when making the call, but…)