Re: [Quantlib-users] Using NaN in Quantlib

[<mat...@gm...>]

Hi David,

 

Because of (c), I assume you are talking about the C++ version of QuantLib.

 

There is, for example, this:  https://github.com/lballabio/QuantLib/blob/60afa352713bac2b759332686ed3070d357cfde8/ql/utilities/null.hpp#L76

Referencing: https://github.com/lballabio/QuantLib/blob/60afa352713bac2b759332686ed3070d357cfde8/ql/qldefines.hpp#L181

 

I see there is also https://github.com/lballabio/QuantLib/blob/60afa352713bac2b759332686ed3070d357cfde8/ql/math/matrixutilities/sparsematrix.hpp#L68

Maybe a similar wrapper for something like vector_sparse.hpp in boost/uBLAS could also work (better) for your use case?

 

Best regards,

Matthias

 

 

From: da...@el... <da...@el...> 
Sent: Monday, 23 November 2020 18:25
To: qua...@li...
Subject: [Quantlib-users] Using NaN in Quantlib

 

Hi, hope everyone is well.

 

A simple question: does QuantLib define NaN somewhere? I have a sparse QuantLib::Array and want to set un-used elements to NaN so I can see which elements are valid and which aren’t.

a) Is there any conceptual problem with this?
b) how do I set a Real to NaN in QL?
c) should I be using a std::vector instead, which could hold a pair<bool,Real> with the bool used to denote a valid value?

 


Thanks

David Sansom