PYTHON-5508 - Add built-in DecimalEncoder and DecimalDecoder #2499
Conversation
bson/decimal128.py
Outdated
| """Converts Python :class:`decimal.Decimal` to BSON :class:`Decimal128`. | ||
|
|
||
| .. warning:: When converting BSON data types to and from built-in data types, | ||
| the possibility of data loss is always present due to mismatches in underlying implementations. |
There was a problem hiding this comment.
"possibility of data loss is always present" is extremely alarming. We have tests that do auto-convert and they don't have data loss issues so I think the language is too strong. Can we identify and explain exactly where the risk is and when it would occur?
There was a problem hiding this comment.
Decimal128 supports up to 34 decimal digits of precision, decimal.Decimal by default supports up to 28 digit during arithmetic operations. When creating a new decimal.Decimal value, that 28 digit limit does not exist, so I don't think we need any warning here.
There was a problem hiding this comment.
Yeah it seems like the truncation issue will be caught on the encoding side:
>>> Decimal128(decimal.Decimal('1'*34))
Decimal128('1111111111111111111111111111111111')
>>> Decimal128(decimal.Decimal('1'*35))
Traceback (most recent call last):
File "<python-input-6>", line 1, in <module>
Decimal128(decimal.Decimal('1'*35))
~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/shane/git/mongo-python-driver/bson/decimal128.py", line 218, in __init__
self.__high, self.__low = _decimal_to_128(value)
~~~~~~~~~~~~~~~^^^^^^^
File "/Users/shane/git/mongo-python-driver/bson/decimal128.py", line 76, in _decimal_to_128
value = ctx.create_decimal(value)
decimal.Inexact: [<class 'decimal.Inexact'>]
doc/changelog.rst
Outdated
| Changes in Version 4.15.0 (XXXX/XX/XX) | ||
| -------------------------------------- | ||
| .. warning:: When converting BSON data types to and from built-in data types, the possibility of data loss is always present | ||
| due to mismatches in underlying implementations. |
There was a problem hiding this comment.
I'm not sure we need this warning here too.
| return Decimal128(value) | ||
|
|
||
|
|
||
| class DecimalDecoder(TypeDecoder): |
There was a problem hiding this comment.
Could we add a docstring example showing out to use these classes?
No description provided.