DOC: list supported float dtypes I #50738
Conversation
doc/source/user_guide/basics.rst
Outdated
| +-------------------------------------------------+---------------------------+--------------------+-------------------------------+----------------------------------------+ | ||
| | :ref:`Boolean (with NA) <api.arrays.bool>` | :class:`BooleanDtype` | :class:`bool` | :class:`arrays.BooleanArray` | ``'boolean'`` | | ||
| +-------------------------------------------------+---------------------------+--------------------+-------------------------------+----------------------------------------+ | ||
| +---------------------------------------------------------------------------------------------+---------------------------+--------------------+-------------------------------+----------------------------------------+ |
There was a problem hiding this comment.
Could you post a screenshot of how this looks? It's somewhat hard to see what changed from the diff
There was a problem hiding this comment.
so, the issue is fixed? Right ! or can i still push my changes? I am new in open source
There was a problem hiding this comment.
Yes, if interested in contributing you can find an issue that hasn't been assigned yet or doesn't have an open pull request
There was a problem hiding this comment.
Ok, thank you. So on an issue only one person can pull a request at one time?
There was a problem hiding this comment.
This is a screenshot of that part of the table.
As Marco said, there is no need to add NumPy integer and NumPy float to the table of pandas extension types. I corrected the table and left only "nullable float" under "nullable integer" (without the rest). This implies that we shouldn’t resize the table.
doc/source/user_guide/basics.rst
Outdated
| | `NumPy integer <https://numpy.org/doc/stable/reference/arrays.scalars.html#numpy.integer>`_ | :class:`numpy.integer` | :class:`int` | :class:`arrays.IntegerArray` | ``'int8'``, ``'int16'``, ``'int32'``, | | ||
| | | | | | ``'int64'``, ``'uint8'``, ``'uint16'``,| | ||
| | | | | | ``'uint32'``, ``'uint64'`` | |
There was a problem hiding this comment.
I don't think this needs adding, as this table says
The following table lists all of pandas extension types
All that needs adding is "nullable float" under "nullable integer"
There was a problem hiding this comment.
Thank you, @MarcoGorelli . As you suggested, I added "nullable float" to the table and removed integer and float.
I noticed that there is no link to “nullable float”
Wouldn’t be useful to have a description and examples for “nullable float”, as we have for “nullable integer”
There was a problem hiding this comment.
yeah, maybe - currently there's no user guide page on nullable floats, but perhaps there should be
pandas/core/tools/numeric.py
Outdated
| - 'float': smallest float dtype (min.: np.float32) | ||
| - 'float': float dtype (min.: np.float32), float16 is not supported. |
There was a problem hiding this comment.
TBH I think this one's clear enough as it is - if it says min: np.float32, then that already implies float16 not being supported
There was a problem hiding this comment.
This is a good point. After re-consideration, I agree.
doc/source/user_guide/basics.rst
Outdated
| | :ref:`Boolean (with NA) <api.arrays.bool>` | :class:`BooleanDtype` | :class:`bool` | :class:`arrays.BooleanArray` | ``'boolean'`` | | ||
| +-------------------------------------------------+---------------------------+--------------------+-------------------------------+----------------------------------------+ | ||
| +--------------------------------------------------+---------------------------+--------------------+-------------------------------+----------------------------------------+ | ||
| | Kind of Data | Data Type | Scalar | Array | String Aliases | |
There was a problem hiding this comment.
looks like this has gone forward by one space 😄 any chance you could bring it back to where it was, to minimise the diff?
There was a problem hiding this comment.
Thank you, @MarcoGorelli . I was wondering why my diff was so unreadable. I deleted one space, now it looks nicer 😄.
Also, I reverted line 73 as you suggested.
pandas/core/tools/numeric.py
Outdated
| - 'float': smallest float dtype (min.: np.float32) | ||
| - 'float': float dtype (min.: np.float32), float16 is not supported. |
|
thx @natmokval |
Updated documentation for
pandas.to_numeric(), pointed out thatdowncastparameter doesn’t supportfloat16dtype. Listed in the description of dtype supported floating point types and added Link to NumPy types.