python - 类型错误 : include and exclude must both be non-string sequences

Question

我有各种数据类型的数据:

1. 餐厅 ID int64
2. 餐厅名称对象
3. 国家/地区代码 int64

...

对文本对象进行评分

投票 int64

数据类型:对象

当我运行下面的代码时:

too_verbose_columns = (data.select_dtypes(include = 'O')\
    .columns[data.select_dtypes(include='O').nunique() > 2]\  
    .tolist())

我收到此错误:

TypeError: include and exclude must both be non-string sequences.

如何解决此错误？

Answer 1

Pandas v0.21 之前的版本

字符串'O'不被视为序列。例如，您需要输入一个列表。此外，无需重复调用select_dtypes:

data = pd.DataFrame([['1', '2', '3'], ['2', '2', '4'], ['5', '6', '7']],
                    columns=['col1', 'col2', 'col3'])

df = data.select_dtypes(include=['O'])

too_verbose_columns = df.columns[df.apply(pd.Series.nunique) > 2].tolist()

print(too_verbose_columns)

['col1', 'col3']

col2 超出范围，因为它只有 2 个唯一元素。

Pandas v0.21+

从 Pandas v0.21 开始，select_dtypes now accepts a scalar :

DataFrame.select_dtypes() now accepts scalar values for include/exclude as well as list-like. (GH16855)