.NET读取Office文件内容(word、excel、ppt)
引用命名空间
1 using Microsoft.Office.Core; 2 using Word = Microsoft.Office.Interop.Word; 3 using Excel = Microsoft.Office.Interop.Excel; 4 using PowerPoint = Microsoft.Office.Interop.PowerPoint;
Word文件的读取
1 public string ReadFile() 2 { 3 string text = string.Empty; 4 Word.ApplicationClass app = null; 5 Word.Document doc = null; 6 object readOnly = true; 7 object missing = System.Reflection.Missing.Value; 8 object fileName = this.FileInstance.FullName; 9 try 10 { 11 app = new Microsoft.Office.Interop.Word.ApplicationClass(); 12 doc = app.Documents.Open(ref fileName, ref missing, ref readOnly, ref missing, ref missing, ref missing, ref missing, ref missing, ref missing, ref missing, ref missing, ref missing, ref missing, ref missing, ref missing, ref missing); 13 text = doc.Content.Text.Replace("\r", string.Empty).Replace("\n", string.Empty).Replace("\t", string.Empty); 14 } 15 catch 16 { 17 18 } 19 finally 20 { 21 doc.Close(ref missing, ref missing, ref missing); 22 doc = null; 23 app.Quit(ref missing, ref missing, ref missing); 24 app = null; 25 } 26 return text; 27 }
Excel文件的读取
1 public string ReadFile() 2 { 3 string text = string.Empty; 4 Excel.ApplicationClass app = null; 5 Excel.Workbook book = null; 6 object readOnly = true; 7 object missing = System.Reflection.Missing.Value; 8 object fileName = this.FileInstance.FullName; 9 try 10 { 11 app = new Microsoft.Office.Interop.Excel.ApplicationClass(); 12 book = app.Workbooks.Open(fileName.ToString(), missing, readOnly, missing, missing, missing, missing, missing, missing, missing, missing, missing, missing, missing, missing); 13 foreach (Excel.Worksheet sheet in book.Sheets) 14 { 15 for (int i = 1; i <= sheet.UsedRange.Cells.Rows.Count; i++) 16 { 17 for (int j = 1; j <= sheet.UsedRange.Cells.Columns.Count; j++) 18 { 19 text += ((Excel.Range)sheet.Cells[i, j]).Text.ToString().Replace("\r", string.Empty).Replace("\n", string.Empty).Replace("\t", string.Empty) + " "; 20 } 21 } 22 } 23 } 24 catch 25 { 26 27 } 28 finally 29 { 30 book.Close(missing, fileName, missing); 31 book = null; 32 app.Quit(); 33 app = null; 34 } 35 return text; 36 }
PPT文件的读取
1 public override string ReadFile() 2 { 3 string text = string.Empty; 4 PowerPoint.ApplicationClass app = null; 5 PowerPoint.Presentation pp = null; 6 object readOnly = true; 7 object missing = System.Reflection.Missing.Value; 8 object fileName = this.FileInstance.FullName; 9 10 try 11 { 12 app = new Microsoft.Office.Interop.PowerPoint.ApplicationClass(); 13 pp = app.Presentations.Open(fileName.ToString(), Microsoft.Office.Core.MsoTriState.msoTrue, Microsoft.Office.Core.MsoTriState.msoFalse, Microsoft.Office.Core.MsoTriState.msoFalse); 14 15 foreach (PowerPoint.Slide slide in pp.Slides) 16 { 17 foreach (PowerPoint.Shape shape in slide.Shapes) 18 { 19 text += shape.TextFrame.TextRange.Text.Replace("\r", string.Empty).Replace("\n", string.Empty).Replace("\t", string.Empty) + " "; 20 } 21 } 22 } 23 catch 24 { 25 26 } 27 finally 28 { 29 pp.Close(); 30 pp = null; 31 app.Quit(); 32 app = null; 33 } 34 35 return text; 36 }