alteryx / alteryx/woodwork

Add UUID LogicalType

未关闭
#1,384 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看
new feature
主要语言
Python
星标
155
派生
24
PR 合并指标
30 天内没有已合并 PR

描述

It would be nice to have a `UUID` logical type that could be used for columns containing UUID data. Currently if a user creates a column that contains UUID objects and initializes Woodwork on this column, the UUID data will be converted to a `string` dtype. If this new type is added, we would need to determine whether to continue to store the data as UUID objects or whether to convert these to strings as we do today, but it might be preferable to store these as UUID objects.

Current behavior:
```python
>>> import uuid
>>> import pandas as pd
>>> import woodwork as ww

>>> df = pd.DataFrame({"id": [uuid.uuid4() for _ in range(10)], "vals": range(10)})
>>> df.dtypes
id object
vals int64
dtype: object
>>> type(df['id'][0])

>>> df.ww.init()
>>> df.dtypes
id string
vals int64
dtype: object
>>> type(df['id'][0])

```

贡献指南

打开贡献指南

评估

这个 Issue 还没有评估数据。

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。