上一篇文章分享了用户下载任务的提交与下载，以及系统中对任务的保存方式，这篇文章，主要分享excel处理部分。excel处理基本流程如下：
excel处理线程首先要获得生成excel表格的基本数据(其中不包括大的查询数据)，保存到新创建的beansMap中；然后对于大的数据多次查询，将多次查询结果List统一保存到一个List中，最后保存到beansMap中，通过jxls生成excel报表。
public void run() {
if(orderViewService != null && pager != null) {
if(beansMap == null)
beansMap = new HashMap<String, Object>();
// save query result list.
List<OrderView> views = new LinkedList<OrderView>();
// temp pager
Pager tmpPager = new Pager();
// copy pager to tmpPager
try {
BeanUtils.copyProperties(tmpPager, pager);
tmpPager.setPage(1);
} catch(InvocationTargetException e) {
e.printStackTrace();
} catch(IllegalAccessException e) {
e.printStackTrace();
}
// query order view.
tmpPager.setPage(1);
tmpPager.setPageSize(SIZE_PER_PAGE);
List<OrderView> orderViews = null;
do {
orderViews = orderViewService.find(tmpPager);
tmpPager.setPage(tmpPager.getPage() + 1);
views.addAll(orderViews);
} while (ObjectUtil.isNotEmpty(orderViews) && orderViews.size() == 10);
// order type
beansMap.put("eshop", 0);
beansMap.put("direct", 1);
beansMap.put("distribution", 2);
// shop list bean
List<Eshop> shopList = eshopService.findAll();
Map<Integer, String> shopMap = new HashMap<Integer, String>();
for (Eshop shop : shopList) {
shopMap.put(shop.getId(), shop.getName());
}
shopMap.put(-1, "-线下-");
shopMap.put(null, "-空-");
// logistic name map bean
Map<Integer, String> logisticMap = new HashMap<Integer, String>();
List<Logistic> logisticList = logisticService.findAll();
for (Logistic logistic : logisticList) {
logisticMap.put(logistic.getId(), logistic.getName());
}
// other beans
beansMap.put("PayMode", Erp.PAYMODEMAP);
beansMap.put("ShippingTypes", Erp.ShippingTypes.map);
beansMap.put("OrderPayTypes", Erp.OrderPayTypes.map);
beansMap.put("OrderStatus", Erp.OrderStatus.statusMap);
beansMap.put("FLAGS", OrderViewServiceImpl.SELLERFLAGS);
// current time.
beansMap.put("CURRENT_TIME", new Date());
beansMap.put("data", views);
// generate excel.
ExportResult exportResult = exportConfigService.createExportFile(jobParams.getTemplateFileCode(), jobParams.getGenerateExcelFilePath(), beansMap);
// update job parameter.
generateExportFile(jobParams, exportResult.getWorkBook());
}
}
/**
* write <code>Workbook</code> object to file.
* @param jobParams job parameters
* @see AsynchronyExportJobParams
* @param workbook excel variable
* @see Workbook
* @return if success return true else return false
*/
private Boolean generateExportFile(AsynchronyExportJobParams jobParams, Workbook workbook) {
// save excel into file.
final String generateFilePrefix = Erp.contextPath + "/export/generate/";
Boolean res = true;
try {
OutputStream os = new BufferedOutputStream(new FileOutputStream(generateFilePrefix + jobParams.getGenerateExcelFilePath()));
workbook.write(os);
os.flush();
os.close();
} catch(IOException e) {
res = false;
}
// update job parameters in DB.
if(res) {
jobParams = asynchronyExportService.load(jobParams.getId());
jobParams.setJobCurrentStatus(JobStatus.FINISHED);
jobParams.setJobFinishTime(new Date());
jobParams.setIsDeleteExportedFile(false);
if(!asynchronyExportService.saveOrUpdate(jobParams))
res = false;
}
return res;
}
对于excel的管理，我采用newSingleThreadExecutor管理，基本Service类如下：
public class ProcessOrderExportService {
private static ExecutorService exec = Executors.newSingleThreadExecutor();
/**
* add a job into executor queue.
*/
public static void pushJob(Runnable job) {
exec.execute(job);
}
/**
* shutdown service.
*/
public static void shutdownService() {
// interrupted all threads.
exec.shutdown();
// wait for two seconds.
try {
TimeUnit.SECONDS.sleep(2);
} catch(InterruptedException e) {
e.printStackTrace();
}
// shutdown all threads now.
if(!exec.isShutdown())
exec.shutdownNow();
}
/**
* restart service.
*/
public static void restartService() {
if(!exec.isTerminated())
exec.shutdownNow();
exec = Executors.newSingleThreadExecutor();
}
}
这样这完成了excel导出任务的处理。对于不同的导出方式，可以创建不同的Runnable线程。线程中常见的问题就是数据的传递方式，可以采用二种方法，第一是采用数据Map<String, Object>在线程的构造方法中传递，例如：public ProcessOrderExportJob(Map<String, Object> paramMap)。第二种，线程中数据传递spring注入的方式，基本采用ApplicationContext.getBean(SomeClass.class)方式获取各种类。当然，与可以两种方式混合使用。我做异步的系统中采用SpringMVC框架，上面的线程参数传递，我采用混合的方式，线程构造方法如下：
/**
* query conditions.
*/
private Pager pager;
/**
* beans map.
*/
private Map<String, Object> beansMap;
/**
* order view service.
*/
private OrderViewService orderViewService;
/**
* asynchrony export job parameters service.
*/
private AsynchronyExportJobParamsService asynchronyExportService;
/**
* export configuration service.
*/
private ExportConfigService exportConfigService;
/**
* eshop service.
*/
private EshopService eshopService;
/**
* logistic service.
*/
private LogisticService logisticService;
/**
* job parameters.
*/
private AsynchronyExportJobParams jobParams;
/**
* the constant variable of a size per page.
*/
private final static int SIZE_PER_PAGE = 10;
public ProcessOrderExportJob(Pager pager, Map<String, Object> beansMap, AsynchronyExportJobParams jobParams) {
this.pager = pager;
this.beansMap = beansMap;
this.jobParams = jobParams;
orderViewService = Erp.context.getBean(OrderViewService.class);
asynchronyExportService = Erp.context.getBean(AsynchronyExportJobParamsService.class);
exportConfigService = Erp.context.getBean(ExportConfigService.class);
eshopService = Erp.context.getBean(EshopService.class);
logisticService = Erp.context.getBean(LogisticService.class);
}
在线程的构造函数中，完成了参数的初始化。