C#语言实现PDF转Excel

• 实现效果

• 第三方库

ClosedXML

iTextSharp

• 实现源码

using System.Text;
using iTextSharp.text.pdf;
using iTextSharp.text.pdf.parser;
using System.Text.RegularExpressions;
using ClosedXML.Excel;

namespace PdfToExcel_winform
{
    public partial class MainForm : Form
    {
        public MainForm()
        {
            InitializeComponent();
        }

        private void Form1_Load(object sender, EventArgs e)
        {
            DateTime specificDate = new DateTime(9999, 3, 15); // 试用截至日期
            DateTime currentDate = DateTime.Now; // 获取当前日期和时间

            if (specificDate < currentDate)
            {
                AddTextAndScroll("******************************************************************************");                
                AddTextAndScroll("==========软件试用已到期，请联系软件开发者授权续期==========");
                AddTextAndScroll("******************************************************************************");
              
                button1.Enabled = false;
            }

        }

        private void textBox1_TextChanged(object sender, EventArgs e)
        {
            AddTextAndScroll("=================程序开始执行，START=================");
            string pdfPath = textBox1.Text;
            using (PdfReader reader = new PdfReader(pdfPath))
            {
                StringBuilder text = new StringBuilder();

                for (int i = 1; i <= reader.NumberOfPages; i++)
                {
                    text.Append(PdfTextExtractor.GetTextFromPage(reader, i));
                }
                //    richTextBox1.Text = text.ToString();
                
                AddTextAndScroll("文件读取完成，原始文件信息：\n" + text.ToString());
                ReadLine(text.ToString(), pdfPath);
            }
        }

        private void button1_Click(object sender, EventArgs e)
        {
            if (openFileDialog1.ShowDialog() == DialogResult.OK)
            {
                // 获取选择的文件路径
                string selectedFilePath = openFileDialog1.FileName;
                // 处理文件路径，例如显示在文本框中或标签中

                textBox1.Text = selectedFilePath;

            }
        }

        private void richTextBox1_TextChanged(object sender, EventArgs e)
        {
            // richTextBox1.ScrollToCaret();
        }


        private void AddTextAndScroll(string text)
        {
            // 添加文本到RichTextBox
            richTextBox1.AppendText(text + Environment.NewLine);
            // 滚动到RichTextBox的最后
            richTextBox1.ScrollToCaret();
        }

        private void ReadLine(string text, string path)
        {
            List<Dictionary<string, string>> dataDictionaryList = new List<Dictionary<string, string>>();

            using (StringReader reader = new StringReader(text.ToString()))
            {
                string line, line1;
                while ((line = reader.ReadLine()) != null)
                {
                    if (line.StartsWith(" ") && line.Length > 5)
                    {
                        try
                        {
                          //  AddTextAndScroll("11111：" + line);
                            line = Regex.Replace(line, @"\s+", " ");
                            string[] split = line.Trim().Split(" ");
                            AddTextAndScroll("按行提取有用文件内容：" + string.Join(" ", split));
                            line1 = reader.ReadLine();
                            line1 = Regex.Replace(line1, @"\s+", " ");
                            string[] split1 = line1.Trim().Split(" ");
                            AddTextAndScroll("按行提取有用文件内容：" + string.Join(" ", split1));
                            string total = split1.Last();
                            split1 = split1.Take(split1.Length - 1).ToArray();
                            Dictionary<string, string> dataDictionary = new Dictionary<string, string>();
                         //   AddTextAndScroll(" 长度=====：" + split.Length);
                            dataDictionary.Add("no", split[0]);
                            dataDictionary.Add("sum", split[1]);
                            dataDictionary.Add("desc", string.Join(" ", split1));
                            dataDictionary.Add("xh", split[3]);
                            dataDictionary.Add("price", split[5]);
                            dataDictionary.Add("total", total);
                            AddTextAndScroll("行数据处理完成" );
                            dataDictionaryList.Add(dataDictionary);
                        }
                        catch (Exception ex)
                        {

                            AddTextAndScroll("程序出错：" + ex.Message);

                        }


                    }
                }

                //调用excel处理逻辑
                WriteExcel(dataDictionaryList, path);
            }
        }

            private void WriteExcel(List<Dictionary<string, string>> dataDictionaryList, string path)
            {

                string fileName = System.IO.Path.GetFileNameWithoutExtension(path);
                string directoryPath = System.IO.Path.GetDirectoryName(path);
                string excelPath = System.IO.Path.Combine(directoryPath, fileName + ".xlsx");
                AddTextAndScroll("Excel文件创建完成，文件路径为：" + excelPath);
                // 创建一个新的Excel文件
                using (var workbook = new XLWorkbook())
                {
                    // 添加一个工作表
                    var worksheet = workbook.AddWorksheet("Sheet1");
                
                // 数据填充
                    worksheet.Cell(1, 1).Value = "序号";
                    worksheet.Cell(1, 2).Value = "数量";
                    worksheet.Cell(1, 3).Value = "描述";
                    worksheet.Cell(1, 4).Value = "型号";
                    worksheet.Cell(1, 5).Value = "单价";
                    worksheet.Cell(1, 6).Value = "金额";
                    var dataList = dataDictionaryList.ToArray();
                    for (var i = 0; i < dataList.Length; i++) {
                    //worksheet.Cell(i + 2, 1).Value = Convert.ToInt16(dataDictionaryList[i]["no"]);
                    //worksheet.Cell(i + 2, 2).Value = Convert.ToDouble(dataDictionaryList[i]["sum"]);
                    //worksheet.Cell(i + 2, 3).Value = dataDictionaryList[i]["desc"];
                    //worksheet.Cell(i + 2, 4).Value = dataDictionaryList[i]["xh"];
                    //worksheet.Cell(i + 2, 5).Value = dataDictionaryList[i]["price"];
                    //worksheet.Cell(i + 2, 6).Value = Convert.ToDecimal(dataDictionaryList[i]["total"]);
                    worksheet.Cell(i + 2, 1).Value = dataDictionaryList[i]["no"];
                    worksheet.Cell(i + 2, 2).Value = dataDictionaryList[i]["sum"];
                    worksheet.Cell(i + 2, 3).Value = dataDictionaryList[i]["desc"];
                    worksheet.Cell(i + 2, 4).Value = dataDictionaryList[i]["xh"];
                    worksheet.Cell(i + 2, 5).Value = dataDictionaryList[i]["price"];
                    worksheet.Cell(i + 2, 6).Value = dataDictionaryList[i]["total"];
                    AddTextAndScroll($"第{i+1}行数据写入Excel表");
                }
                AddTextAndScroll("设置Excel表格列宽自适用");
                worksheet.Columns().AdjustToContents();
                worksheet.Column(1).Width = 7;
                // 保存Excel文件
                workbook.SaveAs(excelPath);
                }


                AddTextAndScroll("开始保存Excel文件");
                AddTextAndScroll("Excel文件保存完成，文件路径为：" + excelPath);
                AddTextAndScroll("=================程序执行结束，END=================");

            }

        }
    }