尝试过很多Excel导入导出方法,都不太理想,无意中逛到oschina时,发现了NPOI,无需Office COM组件且不依赖Office,顿时惊为天人,怀着无比激动的心情写下此文。
曾使用过的方法
- 直接导出html,修改后缀名为.xls,这个方法有点像骗人的把戏,而且不能再导入
- 使用Jet OLEDB引擎来进行导入导出,完全使用sql语句来进行操作,缺点能控制的东西非常有限,比如格式就难以控制
- 使用Office COM组件进行导入导出,对环境依赖性太强(如“检索 COM 类工厂…”错误);且需要通过打开Excel.exe进程进行操作;虽然可以通过关闭工作表以及Marshal.ReleaseComObject方法来释放资源,但依然避免不了性能差。
关于NPOI
NPOI是POI项目的.NET版本,是由@Tony Qu(http://tonyqus.cnblogs.com/)等大侠基于POI开发的,可以从http://npoi.codeplex.com/下载到它的最新版本。它不使用Office COM组件(Microsoft.Office.Interop.XXX.dll),不需要安装Microsoft Office,支持对Office 97-2003的文件格式,功能比较强大。更详细的说明请看作者的博客或官方网站。
它的以下一些特性让我相当喜欢:
- 支持对标准的Excel读写
- 支持对流(Stream)的读写 (而Jet OLEDB和Office COM都只能针对文件)
- 支持大部分Office COM组件的常用功能
- 性能优异 (相对于前面的方法)
- 使用简单,易上手
使用NPOI
本文使用的是它当前的最新版本1.2.4,此版本的程序集缩减至2个:NPOI.dll、Ionic.Zip.dll,直接引用到项目中即可。
对于我们开发者使用的对象主要位于NPOI.HSSF.UserModel空间下,主要有HSSFWorkbook、HSSFSheet、HSSFRow、HSSFCell,对应的接口为位于NPOI.SS.UserModel空间下的IWorkbook、ISheet、IRow、ICell,分别对应Excel文件、工作表、行、列。
简单演示一下创建一个Workbook对象,添加一个工作表,在工作表中添加一行一列:
01 | using NPOI.HSSF.UserModel; |
02 | using NPOI.SS.UserModel; |
08 |
IWorkbook workbook = new HSSFWorkbook(); |
09 |
ISheet sheet = workbook.CreateSheet( "Sheet1" ); |
10 |
IRow row = sheet.CreateRow(0); |
11 |
ICell cell = row.CreateCell(0); |
12 |
cell.SetCellValue( "test" ); |
相应的读取代码:
02 | using NPOI.HSSF.UserModel; |
03 | using NPOI.SS.UserModel; |
07 |
void GetSheet(Stream stream) |
09 |
IWorkbook workbook = new HSSFWorkbook(stream); |
10 |
ISheet sheet = workbook.GetSheetAt(0); |
11 |
IRow row = sheet.GetRow(0); |
12 |
ICell cell = row.GetCell(0); |
13 |
string value = cell.ToString(); |
使用NPOI导出
从DataTable读取内容来创建Workbook对象:
01 | public static MemoryStream RenderToExcel(DataTable table) |
03 |
MemoryStream ms = new MemoryStream(); |
07 |
using (IWorkbook workbook = new HSSFWorkbook()) |
09 |
using (ISheet sheet = workbook.CreateSheet()) |
11 |
IRow headerRow = sheet.CreateRow(0); |
14 |
foreach (DataColumn column in table.Columns) |
15 |
headerRow.CreateCell(column.Ordinal).SetCellValue(column.Caption); |
20 |
foreach (DataRow row in table.Rows) |
22 |
IRow dataRow = sheet.CreateRow(rowIndex); |
24 |
foreach (DataColumn column in table.Columns) |
26 |
dataRow.CreateCell(column.Ordinal).SetCellValue(row[column].ToString()); |
如果看不惯DataTable,那么DataReader也行:
01 | public static MemoryStream RenderToExcel(IDataReader reader) |
03 |
MemoryStream ms = new MemoryStream(); |
07 |
using (IWorkbook workbook = new HSSFWorkbook()) |
09 |
using (ISheet sheet = workbook.CreateSheet()) |
11 |
IRow headerRow = sheet.CreateRow(0); |
12 |
int cellCount = reader.FieldCount; |
15 |
for ( int i = 0; i < cellCount; i++) |
17 |
headerRow.CreateCell(i).SetCellValue(reader.GetName(i)); |
24 |
IRow dataRow = sheet.CreateRow(rowIndex); |
26 |
for ( int i = 0; i < cellCount; i++) |
28 |
dataRow.CreateCell(i).SetCellValue(reader[i].ToString()); |
以上代码把创建的Workbook对象保存到流中,可以通过以下方法输出到浏览器,或是保存到硬盘中:
01 | static void SaveToFile(MemoryStream ms, string fileName) |
03 |
using (FileStream fs = new FileStream(fileName, FileMode.Create, FileAccess.Write)) |
05 |
byte [] data = ms.ToArray(); |
07 |
fs.Write(data, 0, data.Length); |
14 | static void RenderToBrowser(MemoryStream ms, HttpContext context, string fileName) |
16 |
if (context.Request.Browser.Browser == "IE" ) |
17 |
fileName = HttpUtility.UrlEncode(fileName); |
18 |
context.Response.AddHeader( "Content-Disposition" , "attachment;fileName=" + fileName); |
19 |
context.Response.BinaryWrite(ms.ToArray()); |
使用NPOI导入
需要注意的是,sheet.LastRowNum = sheet.PhysicalNumberOfRows - 1,这里可能存在BUG:当没有数据或只有一行数据时sheet.LastRowNum为0,PhysicalNumberOfRows 表现正常。
这里读取流中的Excel来创建Workbook对象,并转换成DataTable:
01 | static DataTable RenderFromExcel(Stream excelFileStream) |
03 |
using (excelFileStream) |
05 |
using (IWorkbook workbook = new HSSFWorkbook(excelFileStream)) |
07 |
using (ISheet sheet = workbook.GetSheetAt(0)) |
09 |
DataTable table = new DataTable(); |
11 |
IRow headerRow = sheet.GetRow(0); |
12 |
int cellCount = headerRow.LastCellNum; |
13 |
int rowCount = sheet.LastRowNum; |
16 |
for ( int i = headerRow.FirstCellNum; i < cellCount; i++) |
18 |
DataColumn column = new DataColumn(headerRow.GetCell(i).StringCellValue); |
19 |
table.Columns.Add(column); |
22 |
for ( int i = (sheet.FirstRowNum + 1); i <= rowCount; i++) |
24 |
IRow row = sheet.GetRow(i); |
25 |
DataRow dataRow = table.NewRow(); |
29 |
for ( int j = row.FirstCellNum; j < cellCount; j++) |
31 |
if (row.GetCell(j) != null ) |
32 |
dataRow[j] = GetCellValue(row.GetCell(j)); |
36 |
table.Rows.Add(dataRow); |
或者是直接生成SQL语句来插入到数据库:
01 | public static int RenderToDb(Stream excelFileStream, string insertSql, DBAction dbAction) |
04 |
using (excelFileStream) |
06 |
using (IWorkbook workbook = new HSSFWorkbook(excelFileStream)) |
08 |
using (ISheet sheet = workbook.GetSheetAt(0)) |
10 |
StringBuilder builder = new StringBuilder(); |
12 |
IRow headerRow = sheet.GetRow(0); |
13 |
int cellCount = headerRow.LastCellNum; |
14 |
int rowCount = sheet.LastRowNum; |
16 |
for ( int i = (sheet.FirstRowNum + 1); i <= rowCount; i++) |
18 |
IRow row = sheet.GetRow(i); |
21 |
builder.Append(insertSql); |
22 |
builder.Append( " values (" ); |
23 |
for ( int j = row.FirstCellNum; j < cellCount; j++) |
25 |
builder.AppendFormat( "'{0}'," , GetCellValue(row.GetCell(j)).Replace( "'" , "''" )); |
27 |
builder.Length = builder.Length - 1; |
31 |
if ((i % 50 == 0 || i == rowCount) && builder.Length > 0) |
34 |
rowAffected += dbAction(builder.ToString()); |
这里的Excel可能没有数据,所以可以加一个方法来检测:
01 | public static bool HasData(Stream excelFileStream) |
03 |
using (excelFileStream) |
05 |
using (IWorkbook workbook = new HSSFWorkbook(excelFileStream)) |
07 |
if (workbook.NumberOfSheets > 0) |
09 |
using (ISheet sheet = workbook.GetSheetAt(0)) |
11 |
return sheet.PhysicalNumberOfRows > 0; |