最近需要需要项目开发需要从excel导入数据到数据库，于是就开始找开源的java操作excel的框架。貌似比较流行的有jxl和poi两个框架。网上有些对这两个框架比较的文章，但都不是最近的。根据项目需要，下面对jxl和poi读写excel的性能做个比较。

jxl：jxl-2.6.12.jar

poi:poi-3.8.jar

JXLTestMain.java

package com.nexusy.excel.jxl;

import java.io.File;

import java.io.IOException;

import java.util.List;

import jxl.Cell;

import jxl.CellType;

import jxl.NumberCell;

import jxl.Sheet;

import jxl.Workbook;

import jxl.read.biff.BiffException;

import jxl.write.Label;

import jxl.write.Number;

import jxl.write.WritableSheet;

import jxl.write.WritableWorkbook;

import jxl.write.WriteException;

import jxl.write.biff.RowsExceededException;

import com.nexusy.excel.Record;

import com.nexusy.excel.TestUtil;

public class JXLTestMain {

private final static String filename = "jxltest.xls";

private final static String[] headers = {"ID", "标题", "价格", "数量", "描述"};

private final static int rows = 65535;

public static void main(String[] args) {

writeExcel();

readExcel();

}

public static void writeExcel() {

try {

Thread.sleep(1000*20);

} catch (InterruptedException e1) {

e1.printStackTrace();

}

try {

WritableWorkbook workbook = Workbook.createWorkbook(new File(filename));

WritableSheet sheet = workbook.createSheet("jxl测试", 0);

for (int i = 0; i < headers.length; i++) {

Label label = new Label(i, 0 , headers[i]);

sheet.addCell(label);

}

List records = TestUtil.getRecords(rows);

long s1 = System.nanoTime();

int c = 1;

for (Record record : records) {

sheet.addCell(new Number(0, c, record.getId()));

sheet.addCell(new Label(1, c, record.getTitle()));

sheet.addCell(new Number(2, c, record.getPrice()));

sheet.addCell(new Number(3, c, record.getQuantity()));

sheet.addCell(new Label(4, c, record.getDesc()));

c++;

}

workbook.write();

workbook.close();

long s2 = System.nanoTime();

System.out.println("jxl write " + rows + " rows to excel:" + (s2-s1));

} catch (IOException e) {

e.printStackTrace();

} catch (RowsExceededException e) {

e.printStackTrace();

} catch (WriteException e) {

e.printStackTrace();

}

}

public static void readExcel() {

try {

Thread.sleep(1000*20);

} catch (InterruptedException e1) {

e1.printStackTrace();

}

try {

long s1 = System.nanoTime();

Workbook workbook = Workbook.getWorkbook(new File(filename));

Sheet sheet = workbook.getSheet(0);

System.out.println(sheet.getName());

for(int i = 0; i < sheet.getRows(); i++){

Cell[] cells = sheet.getRow(i);

for (Cell cell : cells) {

if(cell.getType() == CellType.NUMBER){

System.out.print(((NumberCell)cell).getValue()+" ");

} else if(cell.getType() == CellType.LABEL){

System.out.print(cell.getContents()+" ");

}

}

System.out.println();

}

workbook.close();

long s2 = System.nanoTime();

System.out.println("jxl read " + rows + " rows from excel:" + (s2-s1));

} catch (BiffException e) {

e.printStackTrace();

} catch (IOException e) {

e.printStackTrace();

}

}

}

POITestMain.java

package com.nexusy.excel.poi;

import java.io.FileInputStream;

import java.io.FileNotFoundException;

import java.io.FileOutputStream;

import java.io.IOException;

import java.io.InputStream;

import java.util.List;

import org.apache.poi.hssf.usermodel.HSSFWorkbook;

import org.apache.poi.ss.usermodel.Cell;

import org.apache.poi.ss.usermodel.Row;

import org.apache.poi.ss.usermodel.Sheet;

import org.apache.poi.ss.usermodel.Workbook;

import com.nexusy.excel.Record;

import com.nexusy.excel.TestUtil;

public class POITestMain {

private final static String filename = "poitest.xls";

private final static String[] headers = {"ID", "标题", "价格", "数量", "描述"};

private final static int rows = 65535;

public static void main(String[] args) {

writeExcel();

readExcel();

}

public static void writeExcel() {

try {

Thread.sleep(1000*20);

} catch (InterruptedException e1) {

e1.printStackTrace();

}

Workbook wb = new HSSFWorkbook();

try {

FileOutputStream fileOut = new FileOutputStream(filename);

Sheet sheet = wb.createSheet("poi测试");

Row row = sheet.createRow(0);

for (int i = 0; i < headers.length; i++) {

row.createCell(i).setCellValue(headers[i]);

}

List records = TestUtil.getRecords(rows);

long s1 = System.nanoTime();

int r = 1;

for (Record record : records) {

row = sheet.createRow(r);

row.createCell(0).setCellValue(record.getId());

row.createCell(1).setCellValue(record.getTitle());

row.createCell(2).setCellValue(record.getPrice());

row.createCell(3).setCellValue(record.getQuantity());

row.createCell(4).setCellValue(record.getDesc());

r++;

}

wb.write(fileOut);

fileOut.close();

long s2 = System.nanoTime();

System.out.println("poi write " + rows + " rows to excel:" + (s2-s1));

} catch (FileNotFoundException e) {

e.printStackTrace();

} catch (IOException e) {

e.printStackTrace();

}

}

public static void readExcel() {

try {

Thread.sleep(1000*20);

} catch (InterruptedException e1) {

e1.printStackTrace();

}

try {

long s1 = System.nanoTime();

InputStream inp = new FileInputStream(filename);

Workbook wb = new HSSFWorkbook(inp);

Sheet sheet = wb.getSheetAt(0);

System.out.println(sheet.getSheetName());

for(Row row : sheet){

for(Cell cell : row){

switch (cell.getCellType()) {

case Cell.CELL_TYPE_NUMERIC:

System.out.print(cell.getNumericCellValue() + " ");

break;

case Cell.CELL_TYPE_STRING:

System.out.print(cell.getStringCellValue() + " ");

break;

default:

break;

}

}

System.out.println();

}

inp.close();

long s2 = System.nanoTime();

System.out.println("poi read " + rows + " rows from excel:" + (s2-s1));

} catch (FileNotFoundException e) {

e.printStackTrace();

} catch (IOException e) {

e.printStackTrace();

}

}

}

测试结果如下：

jxl写入65535行数据

jxl读取65535行数据：

poi写入65535行数据：

poi读取65535行数据：

大小: 41.4 KB

大小: 38.6 KB

大小: 40.1 KB

大小: 40.2 KB

描述: 项目代码

下载次数: 469

分享到：

2012-06-06 20:48

浏览 19914

评论

11 楼

zi_wu_xian

2016-09-02

不管用POI还是JXL都是第三的操作excel的组件，都有破坏excel文件格式的可能，还是用PageOffice动态生成excel文件，操作excel文件中的数据比较好，因为是调用的Office自己的VBA接口，所以不会有格式问题。

10 楼

lanhuidong

2014-03-29

weimeili8211 写道

我用poi,jxl都 会内存溢出

9 楼

weimeili8211

2014-03-28

我用poi,jxl都 会内存溢出

8 楼

tony_jingzhou

2013-10-16

tangzililiang 写道

这样比较太片面了哦，如果尝试用POI的事件驱动模式读取我相信内存占用会大大降低，JXL将文件一次性加载到内存，读取大文件是短板(原理可参照xml文件读取方式dom和sax的区别)

采用事件模型，poi读取10M以上的Excel 还是会内存溢出？

7 楼

tangzililiang

2013-07-01

这样比较太片面了哦，如果尝试用POI的事件驱动模式读取我相信内存占用会大大降低，JXL将文件一次性加载到内存，读取大文件是短板(原理可参照xml文件读取方式dom和sax的区别)

6 楼

lanhuidong

2013-05-09

bamboo_prince 写道

好像效率差不多嘛，你推荐用哪种啊？

我自己选择了poi

5 楼

bamboo_prince

2013-05-08

好像效率差不多嘛，你推荐用哪种啊？

4 楼

lanhuidong

2013-05-02

winerlinwen1 写道

我怎么读到2760行就爆了！我用poi操作的！

什么爆了？

3 楼

winerlinwen1

2013-04-28

我怎么读到2760行就爆了！我用poi操作的！

2 楼

lanhuidong

2012-09-10

zhangjian420 写道

这个是什么工具监视的！

jvisualvm，jdk自带的工具，在bin目录中。兄弟可有更好的监视工具推荐？

1 楼

zhangjian420

2012-07-24

这个是什么工具监视的！