我公司是做erp系統(tǒng)的,現(xiàn)在要求做一個(gè)款號(hào)轉(zhuǎn)數(shù)的task,由于每天款號(hào)轉(zhuǎn)數(shù)的數(shù)據(jù)量比較大,而且款號(hào)涉及的一些關(guān)聯(lián)要也要進(jìn)行轉(zhuǎn)數(shù),所以有時(shí)候會(huì)導(dǎo)致out of momery ,要進(jìn)行修改! 原代碼是用hibernate實(shí)現(xiàn)的, list = gzBma00160pDAO.getTransferData(Bma00160p.class,object); if(list!=null){ for(Iterator it = list.iterator();it.hasNext();){ Bma00160p newBma00160p = new Bma00160p(); BeanUtils.copyProperties(it.next(), newBma00160p); try{ Bma00160p bma00160p = bma00160pDAO.getBma00160p(newBma00160p.getBma00160pPK()); if(bma00160p!=null){ newBma00160p.setState(BaseObject.UPDATEMODE_MODIFY); }else{ newBma00160p.setState(BaseObject.UPDATEMODE_INSERT); } newBma00160p.setDlastmodifytime(nowtime); bma00160pDAO.saveObject(newBma00160p); }catch(Exception e){ ERPTaskErrorMailListener.addMessage("BMA00160P出錯(cuò),代號(hào): " + newBma00160p.getBma00160pPK()); LogUtils.warn("bma00160p error : " + newBma00160p); e.printStackTrace(); } } list.clear(); } 每個(gè)表查詢轉(zhuǎn)數(shù)的代碼如上,每次都通過本地的對(duì)象get遠(yuǎn)程的對(duì)象,看有沒有存在,有則update否則insert,這樣如果數(shù)據(jù)量大,頻繁的操作數(shù)據(jù)庫(kù)是不是導(dǎo)致性能低下的原因呢?如果我換成jdbc,寫sql語(yǔ)句一次性u(píng)pdate 或 insert 會(huì)不會(huì)性能有所提升?但這樣的話就要建立臨時(shí)表去存放本地的數(shù)據(jù)給遠(yuǎn)程用... 望高手能給予解答,謝謝