如何解决 Pandas 的属性错误：\"AttributeError: module \'pandas\' has no attribute \'Float64Dtype\"？

Question

我正在尝试读取 .sav 文件。为此，我安装了 pyreadstat 使用

conda install -c conda-forge pyreadstat

但是当我运行以下代码行时（甚至在读取文件之前）：

import pyreadstat

我收到此错误：“AttributeError：模块'pandas'没有属性'Float64Dtype'”。

关于如何解决它的任何想法？

完整错误如下

AttributeError                            Traceback (most recent call last)
Input In [3], in <cell line: 2>()
      1 import pandas
----> 2 import pyreadstat

File ~\anaconda3\envs\dhdsblend2021\lib\site-packages\pyreadstat\__init__.py:17, in <module>
      1 # #############################################################################
      2 # Copyright 2018 Hoffmann-La Roche
      3 #
   (...)
     14 # limitations under the License.
     15 # #############################################################################
---> 17 from .pyreadstat import read_sas7bdat, read_xport, read_dta, read_sav, read_por, read_sas7bcat
     18 from .pyreadstat import write_sav, write_dta, write_xport, write_por
     19 from .pyreadstat import set_value_labels, set_catalog_to_sas

File pyreadstat\pyreadstat.pyx:1, in init pyreadstat.pyreadstat()

File pyreadstat\_readstat_writer.pyx:42, in init pyreadstat._readstat_writer()

File ~\anaconda3\envs\dhdsblend2021\lib\site-packages\pandas\__init__.py:258, in __getattr__(name)
    254     from pandas.core.arrays.sparse import SparseArray as _SparseArray
    256     return _SparseArray
--> 258 raise AttributeError(f"module 'pandas' has no attribute '{name}'")

Answer 1

Pyreadstat 至少需要 Pandas 1.2.0。请尝试将 pandas 更新到较新的版本。