docx始终以文本(字符串)的形式从Word表中读取数据.

docx always reads data from Word tables as text (strings).

如果我们要解析具有正确dtypes的数据，则可以执行以下操作之一:

If we want to parse data with correct dtypes we can do one of the following:

• 为所有列手动指定dtype(不灵活)
• 编写我们自己的代码来猜测正确的dtypes(太难了，熊猫IO方法做得很好)
• 将数据转换为CSV格式，并让pd.read_csv()猜测/推断正确的dtypes(我已经选择了这种方式)

• manually specify dtype for all columns (not flexible)
• write our own code to guess correct dtypes (too difficult and , Pandas IO methods do it well)
• convert data into CSV format and let pd.read_csv() guess/infer correct dtypes (I've chosen this way)

非常感谢 @Anton vBR 改进了功能！

Many thanks to @Anton vBR for improving the function!

import pandas as pd
import io
import csv
from docx import Document

def read_docx_tables(filename, tab_id=None, **kwargs):
    """
    parse table(s) from a Word Document (.docx) into Pandas DataFrame(s)

    Parameters:
        filename:   file name of a Word Document

        tab_id:     parse a single table with the index: [tab_id] (counting from 0).
                    When [None] - return a list of DataFrames (parse all tables)

        kwargs:     arguments to pass to `pd.read_csv()` function

    Return: a single DataFrame if tab_id != None or a list of DataFrames otherwise
    """
    def read_docx_tab(tab, **kwargs):
        vf = io.StringIO()
        writer = csv.writer(vf)
        for row in tab.rows:
            writer.writerow(cell.text for cell in row.cells)
        vf.seek(0)
        return pd.read_csv(vf, **kwargs)

    doc = Document(filename)
    if tab_id is None:
        return [read_docx_tab(tab, **kwargs) for tab in doc.tables]
    else:
        try:
            return read_docx_tab(doc.tables[tab_id], **kwargs)
        except IndexError:
            print('Error: specified [tab_id]: {}  does not exist.'.format(tab_id))
            raise

注意:您可能想添加更多检查和异常捕获...

NOTE: you may want to add more checks and exception catching...

示例:

In [209]: dfs = read_docx_tables(fn)

In [210]: dfs[0]
Out[210]:
   A   B               C,X
0  1  B1                C1
1  2  B2                C2
2  3  B3  val1, val2, val3

In [211]: dfs[0].dtypes
Out[211]:
A       int64
B      object
C,X    object
dtype: object

In [212]: dfs[0].columns
Out[212]: Index(['A', 'B', 'C,X'], dtype='object')

In [213]: dfs[1]
Out[213]:
   C1  C2          C3    Text column
0  11  21         NaN  Test "quotes"
1  12  23  2017-12-31            NaN

In [214]: dfs[1].dtypes
Out[214]:
C1              int64
C2              int64
C3             object
Text column    object
dtype: object

In [215]: dfs[1].columns
Out[215]: Index(['C1', 'C2', 'C3', 'Text column'], dtype='object')

解析日期:

In [216]: df = read_docx_tables(fn, tab_id=1, parse_dates=['C3'])

In [217]: df
Out[217]:
   C1  C2         C3    Text column
0  11  21        NaT  Test "quotes"
1  12  23 2017-12-31            NaN

In [218]: df.dtypes
Out[218]:
C1                      int64
C2                      int64
C3             datetime64[ns]
Text column            object
dtype: object