Support function isNaN

[exmy]

Use case

• support function isNaN

There is a slight difference between the isNaN function in Spark and in CH.

For example, when running the following query:

select isNaN(cast('a' as double));
In Spark: false
In CH: Cannot parse string 'a' as Float64: syntax error at begin of string. Note: there are toFloat64OrZero and toFloat64OrNull functions, which returns zero/NULL instead of throwing exception

So we can replace the cast function with toFloat64OrZero, which will return zero instead of throwing an error for invalid input values.