closes bpo-28955: Clarified comparisons between NaN and number in ref… · python/cpython@ad8a000 (original) (raw)

`@@ -1336,12 +1336,11 @@ built-in types.

1336

` involved, they compare mathematically (algorithmically) correct without loss

1337

` of precision.

1338

1339

`` -

The not-a-number values :const:float('NaN') and :const:Decimal('NaN')

1340


 are special. They are identical to themselves (``x is x`` is true) but

1341


 are not equal to themselves (``x == x`` is false). Additionally,

1342

comparing any number to a not-a-number value

1343


 will return ``False``. For example, both ``3 < float('NaN')`` and

1344


 ``float('NaN') < 3`` will return ``False``.

1339


 The not-a-number values ``float('NaN')`` and ``decimal.Decimal('NaN')`` are

1340

special. Any ordered comparison of a number to a not-a-number value is false.

1341

A counter-intuitive implication is that not-a-number values are not equal to

1342


 themselves. For example, if ``x = float('NaN')``, ``3 < x``, ``x < 3``, ``x

1343


 == x``, ``x != x`` are all false. This behavior is compliant with IEEE 754.

1345

1344

1346

1345

`` * Binary sequences (instances of :class:bytes or :class:bytearray) can be

1347

1346

` compared within and across their types. They compare lexicographically using