Python如何提取Word文档中的表格并导出为Excel

pip install Spire.Doc
pip install Spire.XLS

from spire.doc import *
from spire.doc.common import *
from spire.xls import *
from spire.xls.common import *
import os

# 创建 Word 文档对象并加载文件
document = Document()
document.LoadFromFile("E:/input/项目进度.docx")

# 创建 Excel 工作簿并清空默认工作表
workbook = Workbook()
workbook.Worksheets.Clear()

sheet_index = 0
for s in range(document.Sections.Count):
    section = document.Sections.get_Item(s)
    tables = section.Tables

    for t in range(tables.Count):
        table = tables.get_Item(t)
        # 动态创建工作表并递增索引
        sheet = workbook.Worksheets.Add(f"Table_{sheet_index + 1}")
        # ... (后续数据提取)
        sheet_index += 1

# 遍历行与列
for r in range(table.Rows.Count):
    row = table.Rows.get_Item(r)
    for c in range(row.Cells.Count):
        cell = row.Cells.get_Item(c)
        
        cell_text = ""
        for p in range(cell.Paragraphs.Count):
            paragraph = cell.Paragraphs.get_Item(p)
            cell_text += paragraph.Text.strip() + "\n"

        # 写入 Excel 坐标并设置换行
        sheet.Range[r + 1, c + 1].Text = cell_text.strip()
        sheet.Range[r + 1, c + 1].Style.WrapText = True

from spire.doc import *
from spire.doc.common import *
from spire.xls import *
from spire.xls.common import *
import os

# 创建 Word 文档对象并加载文件
document = Document()
document.LoadFromFile("/input/项目进度.docx")

# 创建 Excel 工作簿并删除默认工作表
workbook = Workbook()
workbook.Worksheets.Clear()

# 遍历 Word 文档中的所有节
sheet_index = 0

for s in range(document.Sections.Count):
    section = document.Sections.get_Item(s)

    # 获取当前节中的所有表格
    tables = section.Tables

    for t in range(tables.Count):

        # 获取当前表格
        table = tables.get_Item(t)
        # 新建工作表
        sheet = workbook.Worksheets.Add(f"Table_{sheet_index + 1}")

        # 遍历表格行
        for r in range(table.Rows.Count):
            row = table.Rows.get_Item(r)

            # 遍历单元格
            for c in range(row.Cells.Count):
                cell = row.Cells.get_Item(c)

                # 提取单元格中的文本
                cell_text = ""
                for p in range(cell.Paragraphs.Count):
                    paragraph = cell.Paragraphs.get_Item(p)
                    cell_text += paragraph.Text.strip() + "\n"

                # 去除末尾换行
                cell_text = cell_text.strip()

                # 写入 Excel 单元格
                sheet.Range[r + 1, c + 1].Text = cell_text
                # 设置自动换行
                sheet.Range[r + 1, c + 1].Style.WrapText = True

        # 自动调整列宽行高
        sheet.AllocatedRange.AutoFitColumns()
        sheet.AllocatedRange.AutoFitRows()
        sheet_index += 1

# 保存为 Excel 文件
workbook.SaveToFile("/output/word表格.xlsx", ExcelVersion.Version2016)
document.Close()
workbook.Dispose()