using Spire.Doc;
using Spire.Doc.Collections;
using Spire.Doc.Interface;
using System.IO;
using System.Text;
namespace ExtractWordTable
{
    internal class Program
    {
        static void Main(string[] args)
        {
            // 創(chuàng)建文檔對象
            Document doc = new Document();
            // 加載Word文檔
            doc.LoadFromFile("表格.docx");
            // 遍歷文檔中的所有節(jié)
            for (int sectionIndex = 0; sectionIndex < doc.Sections.Count; sectionIndex++)
            {
                Section section = doc.Sections[sectionIndex];
                // 獲取當前節(jié)中的所有表格
                TableCollection tables = section.Tables;
                // 遍歷當前節(jié)中的所有表格
                for (int tableIndex = 0; tableIndex < tables.Count; tableIndex++)
                {
                    ITable table = tables[tableIndex];
                    // 用于存儲當前表格的所有數(shù)據(jù)
                    string tableData = "";
                    // 遍歷表格中的所有行
                    for (int rowIndex = 0; rowIndex < table.Rows.Count; rowIndex++)
                    {
                        TableRow row = table.Rows[rowIndex];
                        // 遍歷行中的所有單元格
                        for (int cellIndex = 0; cellIndex < row.Cells.Count; cellIndex++)
                        {
                            TableCell cell = row.Cells[cellIndex];
                            // 提取單元格文本（單元格可能包含多個段落）
                            string cellText = "";
                            for (int paraIndex = 0; paraIndex < cell.Paragraphs.Count; paraIndex++)
                            {
                                cellText += (cell.Paragraphs[paraIndex].Text.Trim() + " ");
                            }
                            // 拼接單元格文本，用制表符分隔不同單元格
                            tableData += cellText.Trim();
                            if (cellIndex < row.Cells.Count - 1)
                            {
                                tableData += "\t";
                            }
                        }
                        // 行結束后換行
                        tableData += "\n";
                    }
                    // 保存表格數(shù)據(jù)到文本文件）
                    string filePath = Path.Combine("Tables", $"Section{sectionIndex + 1}_Table{tableIndex + 1}.txt");
                    File.WriteAllText(filePath, tableData, Encoding.UTF8);
                }
            }
            doc.Close();
        }
    }
}

using Spire.Xls;
// 將 tableData 的二維數(shù)組寫入 Workbook

string[] files = Directory.GetFiles(@"C:\Docs", "*.docx");
foreach (string file in files)
{
    doc.LoadFromFile(file);
    // ... 提取邏輯
}

Install-Package DocumentFormat.OpenXml

using DocumentFormat.OpenXml.Packaging;
using DocumentFormat.OpenXml.Wordprocessing;
using System;
using System.Collections.Generic;
public class TableData
{
    public string CellText { get; set; }
    // 可根據(jù)需要擴展行列信息
}
public static List<List<string>> ExtractTablesFromWord(string filePath)
{
    var allTables = new List<List<string>>();
    using (WordprocessingDocument doc = WordprocessingDocument.Open(filePath, false))
    {
        // 獲取文檔主體
        Body body = doc.MainDocumentPart.Document.Body;
        // 遍歷所有表格
        IEnumerable<DocumentFormat.OpenXml.Wordprocessing.Table> tables = body.Elements<DocumentFormat.OpenXml.Wordprocessing.Table>();
        foreach (var table in tables)
        {
            var tableData = new List<string>();
            // 遍歷所有行
            foreach (TableRow row in table.Elements<TableRow>())
            {
                // 遍歷所有單元格
                foreach (TableCell cell in row.Elements<TableCell>())
                {
                    // 獲取單元格文本
                    string cellText = cell.InnerText;
                    tableData.Add(cellText);
                }
            }
            allTables.Add(tableData);
        }
    }
    return allTables;
}

Install-Package Spire.Doc

using Spire.Doc;
using Spire.Doc.Collections;
using System.Text;
public static void ExtractTablesWithSpire(string docPath, string outputPath)
{
    Document doc = new Document();
    doc.LoadFromFile(docPath);
    using (StreamWriter writer = new StreamWriter(outputPath))
    {
        // 遍歷文檔的所有節(jié)
        for (int sectionIdx = 0; sectionIdx < doc.Sections.Count; sectionIdx++)
        {
            Section section = doc.Sections[sectionIdx];
            TableCollection tables = section.Tables;
            // 遍歷每個表格
            for (int tableIdx = 0; tableIdx < tables.Count; tableIdx++)
            {
                ITable table = tables[tableIdx] as ITable;
                writer.WriteLine($"--- 表格 {tableIdx + 1} ---");
                // 遍歷每一行
                for (int rowIdx = 0; rowIdx < table.Rows.Count; rowIdx++)
                {
                    TableRow row = table.Rows[rowIdx];
                    StringBuilder rowText = new StringBuilder();
                    // 遍歷每一列的單元格
                    for (int cellIdx = 0; cellIdx < row.Cells.Count; cellIdx++)
                    {
                        TableCell cell = row.Cells[cellIdx];
                        // 獲取單元格文本
                        string cellText = "";
                        for (int pIdx = 0; pIdx < cell.Paragraphs.Count; pIdx++)
                        {
                            cellText += cell.Paragraphs[pIdx].Text;
                        }
                        rowText.Append(cellText).Append("\t");
                    }
                    writer.WriteLine(rowText.ToString().TrimEnd('\t'));
                }
            }
        }
    }
}

Install-Package GroupDocs.Parser

using GroupDocs.Parser;
using GroupDocs.Parser.Data;
using System;
using System.Collections.Generic;
public static void ExtractTablesWithGroupDocs(string filePath)
{
    using (Parser parser = new Parser(filePath))
    {
        // 檢查文檔是否支持表格提取
        if (!parser.Features.Tables)
        {
            Console.WriteLine("此文檔類型不支持表格提取。");
            return;
        }
        // 獲取表格數(shù)據(jù)
        IEnumerable<PageTableArea> tables = parser.GetTables();
        foreach (PageTableArea table in tables)
        {
            Console.WriteLine($"表格位置: 第 {table.Page.Index} 頁");
            // 遍歷所有行
            for (int row = 0; row < table.RowCount; row++)
            {
                for (int col = 0; col < table.ColumnCount; col++)
                {
                    PageTableAreaCell cell = table[row, col];
                    if (cell != null)
                    {
                        Console.Write(cell.Text?.Trim() ?? "");
                    }
                    Console.Write("\t");
                }
                Console.WriteLine();
            }
            Console.WriteLine("---");
        }
    }
}