python十进制奇怪句柄以及如何获得最接近的十进制浮点数？

Question

The strange result is

In [46]: Decimal(1.1).quantize(Decimal('.1'), rounding=ROUND_UP)
Out[46]: Decimal('1.2')

In [47]: Decimal(1.1).quantize(Decimal('.1'), rounding=ROUND_HALF_UP)
Out[47]: Decimal('1.1')

In [48]: Decimal(3.65).quantize(Decimal('.1'), rounding=ROUND_UP)
Out[48]: Decimal('3.7')

In [49]: Decimal(3.65).quantize(Decimal('.1'), rounding=ROUND_HALF_UP)
Out[49]: Decimal('3.6')

But I want this:

In [47]: Decimal(1.1).quantize(Decimal('.1'), rounding=Somthing)
Out[47]: Decimal('1.1')

In [48]: Decimal(3.65).quantize(Decimal('.1'), rounding=Somthing)
Out[48]: Decimal('3.7')

Which mean I want to get closest float.

Answer 1

The problem is you are feeding float values to Decimal , which is subject to floating point precision errors. Feeding a string removes the problem:

print(Decimal(str(1.1)).quantize(Decimal('.1'), rounding=ROUND_UP))

Decimal('1.1')

print(Decimal(str(3.65)).quantize(Decimal('.1'), rounding=ROUND_UP))

Decimal('3.7')

Specifically, we find:

Decimal(1.1) == Decimal('1.100000000000000088817841970012523233890533447265625')