今天现场说系统中的一个导出Excel的功能导出的文件为空,一测试还真是的,不过奇怪的是导入内容较少的时候还是可以导出的,在内容较多时就为空了。
1. 分析为空情况
A.查询结果为空
B.在struts获取InputStream时为空
看过程序后A情况就被排除了,因为程序中是现在服务器目录生成一个Excel临时文件,然后导出时直接读取的文件流。
然后考虑是不是参数名没和struts的匹配上,一开始还真没找到那个参数,后来才发现只需要有个getXXX方法就行了,参数名就是get后边的名称(应该是默认首字母小写)。
贴上配置文件:
<!-- 增加一个返回结果类型 这个结果类型可以处理用户点击取消下载的时候正确的关闭流 -->
<result-types>
<result-type name="streamx" class="com.sunspoter.lib.web.struts2.dispatcher.StreamResultX"/>
</result-types>
<action name="downloadResult" class="workflowAction">
<result name="success" type="streamx">
<param name="contentType">text/plain</param> <!-- application/octet-stream 无限制类型 -->
<param name="contentDisposition">attachment;fileName="${path}"</param>
<param name="inputName">printResult</param>
<param name="bufferSize">1024*10</param>
</result>
贴上主要方法
public InputStream getPrintResult()
{
System.out.println("getPrintResult");
System.out.println(this.queryStr);
InputStream is = null;
UUID uuid = UUID.randomUUID();
ServletContext context = ServletActionContext.getServletContext();
try
{
this.path = new String(this.path.getBytes(), "ISO8859-1");
}
catch (UnsupportedEncodingException e)
{
e.printStackTrace();
}
String path = context.getRealPath("/") + "case/" + uuid.toString() + ".xls";
System.out.println(path);
try
{
this.queryStr = URLDecoder.decode(this.queryStr, "utf-8");
System.out.println("queryStr=" + this.queryStr);
IWFService wfService = this.workFlowService.getWFService();
DataTableDto dataTableDto = DataTableUtil.getQueryStruct(Integer.parseInt(this.queryid), wfService, null, null, this.queryStr);
if ((dataTableDto != null) && (dataTableDto.Rows != null) &&
(dataTableDto.Rows.size() > 0))
{
List<String> columnList = new ArrayList();
List<ArrayList<Object>> rowList = new ArrayList();
for (DataColumnDto column : dataTableDto.Columns) {
columnList.add(column.getColumnName());
}
for (DataRowDto row : dataTableDto.Rows) {
rowList.add(row.alData);
}
HSSFWorkbook workbook = new HSSFWorkbook();
HSSFSheet sheet = workbook.createSheet("查询结果页");
HSSFRow headRow = sheet.createRow(0);
HSSFCellStyle style = workbook.createCellStyle();
style.setAlignment(HSSFCellStyle.ALIGN_CENTER);
int headRowIndex = 0;
for(DataColumnDto column : dataTableDto.Columns) {
HSSFCell headRowCell = headRow.createCell(headRowIndex++);
headRowCell.setCellValue(column.getColumnName());
headRowCell.setCellStyle(style);
}
int rowIndex = 1;
Iterator localIterator3;
for(DataRowDto dataRow : dataTableDto.Rows) {
HSSFRow row = sheet.createRow(rowIndex++);
int cellIndex = 0;
for(Object item : dataRow.alData) {
HSSFCell cell = row.createCell(cellIndex++);
if(item != null) {
cell.setCellValue(item.toString());
} else {
cell.setCellValue("");
}
cell.setCellStyle(style);
}
}
FileOutputStream fout = new FileOutputStream(path);
workbook.write(fout);
fout.close();
is = context.getResourceAsStream("/case/" + uuid.toString() + ".xls");
return is;
}
}
catch (Exception er)
{
er.printStackTrace();
}
finally
{
try
{
is.close();
File file = new File(path);
if (file.exists())
{
file.delete();
System.out.println("删除文件成功!");
}
}
catch (Exception er)
{
er.printStackTrace();
}
}
return null;
}
在debuge时奇怪的事情发生了
is = context.getResourceAsStream("/case/" + uuid.toString() + ".xls");
没错就是这句代码,大家应该知道这是获取文件流的方法。
但是奇怪的是 当文件比较小时返回的类型为bytearrayInputstream
当文件比较大时返回的类型为FileinputStream
想必到这大家就能发现为什么有时能导出有时导出空白了。被关闭了
这是ByteArrayInputStream
的close()
的方法
/**
* Closing a <tt>ByteArrayInputStream</tt> has no effect. The methods in
* this class can be called after the stream has been closed without
* generating an <tt>IOException</tt>.
* <p>
*/
public void close() throws IOException {
}
可以发现这个方法是空的,也就是说关闭操作对他并没有影响
再看看FileInputStream
方法
public void close() throws IOException {
if (channel != null)
channel.close();
close0();
}
它是可以关闭的,这也是为啥他为空了
再来看看struts实现下载的StreamResultX
类
贴上主要代码
protected void doExecute(String finalLocation, ActionInvocation invocation) throws Exception {
this.resolveParamsFromStack(invocation.getStack(), invocation);
ServletOutputStream oOutput = null;
try {
if(this.inputStream == null) {
this.inputStream = (InputStream)invocation.getStack().findValue(this.conditionalParse(this.inputName, invocation));
}
if(this.inputStream == null) {
String oResponse1 = "StreamResultX : Can not find a java.io.InputStream with the name [" + this.inputName + "] in the invocation stack. " + "Check the <param name=\"inputName\"> tag specified for this action.";
LOG.error(oResponse1, new String[0]);
throw new IllegalArgumentException(oResponse1);
}
HttpServletResponse oResponse = (HttpServletResponse)invocation.getInvocationContext().get("com.opensymphony.xwork2.dispatcher.HttpServletResponse");
if(this.contentCharSet != null && !this.contentCharSet.equals("")) {
oResponse.setContentType(this.conditionalParse(this.contentType, invocation) + ";charset=" + this.contentCharSet);
} else {
oResponse.setContentType(this.conditionalParse(this.contentType, invocation));
}
int iSize1;
if(this.contentLength != null) {
String oBuff = this.conditionalParse(this.contentLength, invocation);
boolean iSize = true;
try {
iSize1 = Integer.parseInt(oBuff);
if(iSize1 >= 0) {
oResponse.setContentLength(iSize1);
}
} catch (NumberFormatException var22) {
LOG.warn("StreamResultX warn : failed to recongnize " + oBuff + " as a number, contentLength header will not be set", var22, new String[0]);
}
}
if(this.contentDisposition != null) {
oResponse.addHeader("Content-Disposition", this.conditionalParse(this.contentDisposition, invocation));
}
if(!this.allowCaching) {
oResponse.addHeader("Pragma", "no-cache");
oResponse.addHeader("Cache-Control", "no-cache");
}
oOutput = oResponse.getOutputStream();
if(LOG.isDebugEnabled()) {
LOG.debug("StreamResultX : Streaming result [" + this.inputName + "] type=[" + this.contentType + "] length=[" + this.contentLength + "] content-disposition=[" + this.contentDisposition + "] charset=[" + this.contentCharSet + "]", new String[0]);
}
byte[] oBuff1 = new byte[this.bufferSize];
try {
LOG.debug("StreamResultX : Streaming to output buffer +++ START +++", new String[0]);
while(-1 != (iSize1 = this.inputStream.read(oBuff1))) {
oOutput.write(oBuff1, 0, iSize1);
}
LOG.debug("StreamResultX : Streaming to output buffer +++ END +++", new String[0]);
oOutput.flush();
} catch (Exception var23) {
LOG.warn("StreamResultX Warn : socket write error", new String[0]);
if(oOutput != null) {
try {
oOutput.close();
} catch (Exception var21) {
oOutput = null;
}
}
} finally {
if(this.inputStream != null) {
this.inputStream.close();
}
if(oOutput != null) {
oOutput.close();
}
}
} finally {
if(this.inputStream != null) {
this.inputStream.close();
}
if(oOutput != null) {
oOutput.close();
}
}
}
可以发现和我们自己写的也没啥区别
一般都这么写
fin = new FileInputStream(file);
response.reset();
response.setCharacterEncoding("utf-8");
response.setContentType("application/msword");
response.setHeader("Content-Disposition", "attachment;filename=" + fileName);
out = response.getOutputStream();
byte[] buffer = new byte[512];
int byteToRead = -1;
while ((byteToRead = fin.read(buffer)) != -1) {
out.write(buffer, 0, byteToRead);
}
out.flush();
但是惊奇的发现它的finally
方法中有对流的关闭方法,好了这就意味着我们在获取完输出流后就不要急着关闭了,struts用完之后就帮我们关闭的。
但是主要问题还是getResourceAsStream()
为啥返回的类型还是可变的……