带有POI和FileOutputStream的java Swing应用程序性能问题

Question

我正在维护swing应用程序。应用程序从access数据库中获取数据并将其填充到excel文件中。 在100条记录之前，当填充超过200条记录时，它可以正常工作。当我看到task manager内存消耗降低时，它会增加到700 mbs。 之前的程序员使用的是正确关闭的Apache poi和FileOutputStrean。 我能理解内存泄漏问题在哪里。我的代码片段是

                final ArrayList<LinkedHashMap<String, Object>> arrlist_b1 = new Model()
              .getAllowanceDetails1(
                currency,
                type,
                arrListYear,
                arrList_survey_name,
                arrList_country,
                arrList_company);

            final DateFormat dateFormat = new SimpleDateFormat(
              "yyyy-MM-dd_HH_mm_ss");
            final Date date = new Date();
            XSSFWorkbook workbook1 = new XSSFWorkbook();

            FileOutputStream out1 = null;
            try {
             out1 = new FileOutputStream(
               new File(
                 filepath));
            } catch (FileNotFoundException e2) {
             e2.printStackTrace();
            }
            XSSFSheet spreadsheet1 = workbook1
              .createSheet("Sheet Name");

            XSSFRow row1 = spreadsheet1
              .createRow(0);
            CellStyle style = workbook1
              .createCellStyle();

            style = new Style()
              .a_b_and_p_heading(workbook1);
            int count_column1 = 0;
            int count_row = 1;

            pb.setMaximum(arrlist_b1
              .size() - 1);
            int total=arrlist_b1
              .size();
            for (int a = 0; a < arrlist_b1
              .size(); a++) {
             pb.setValue(a);

             download_status
               .setText(a
                 + 1
                 + " / "
                 + total);
             System.out
               .println(a);

             LinkedHashMap<String, Object> tmpData = (LinkedHashMap<String, Object>) arrlist_b1
               .get(a);
             Set<String> key = tmpData
               .keySet();

             Iterator it = key
               .iterator();
             String hmData, created_by;
             XSSFRow row_tbl_heading = spreadsheet1
               .createRow(count_row);

             int count_column = 0;
             while (it
               .hasNext()) {
              String hmKey = (String) it
                .next();

              if (a == 0) {
               spreadsheet1
                 .autoSizeColumn(count_column1);
               XSSFCell cell_row = row1
                 .createCell(count_column1);
               cell_row.setCellStyle(style);
               hmKey = hmKey
                 .replaceAll(
                   "_",
                   " ")
                 .toLowerCase();
               hmKey = StringUtils
                 .capitalize(hmKey);

               cell_row.setCellValue(hmKey);
               count_column1++;

              }
              DecimalFormat formatter = new DecimalFormat(
                "##,###");
              XSSFCell cell_row = row_tbl_heading
                .createCell(count_column);
              spreadsheet1
                .autoSizeColumn(count_column);

              CellStyle cellStyle = workbook1
                .createCellStyle();
              cellStyle
                .setDataFormat(workbook1
                  .getCreationHelper()
                  .createDataFormat()
                  .getFormat(
                    "##,###"));

              Object obj = tmpData
                .get(hmKey);
              if (obj instanceof Integer) {

               if ((Integer) tmpData
                 .get(hmKey) == -1) {
                cell_row.setCellValue("");
               } else {
                cell_row.setCellValue((Integer) tmpData
                  .get(hmKey));

               }

              } else if (obj instanceof Double) {

               if (Double
                 .parseDouble(tmpData
                   .get(hmKey)
                   .toString()) == -1.0) {
                cell_row.setCellValue("");
               } else {

                int round_val = new Validation()
                  .count_digit(tmpData
                    .get(hmKey)
                    .toString());
                cell_row.setCellFormula(("ROUND("
                  + Double.parseDouble(tmpData
                    .get(hmKey)
                    .toString())
                  + ",-"
                  + round_val + ")"));
                cell_row.setCellStyle(cellStyle);

               }

              } else {
               hmData = (String) tmpData
                 .get(hmKey);
               cell_row.setCellValue(hmData);
              }
              count_column++;
             }

             count_row++;
             arrlist_b1.remove(a);
             tmpData.remove(key);
             it.remove();

            }
            try {
             workbook1
               .write(out1);

            } catch (Exception e1) {
             System.out
               .println("hello+"
                 + e1.getMessage());
             e1.printStackTrace();
            }
            try {
             out1.close();
            } catch (Exception e1) {
             // TODO
             // Auto-generated
             // catch
             // block
             System.out
               .println("hello+"
                 + e1.getMessage());
             e1.printStackTrace();
            }

400条记录之后，申请没有回复。 我使用Jprofiler跟踪内存使用情况，其中char[]和String消耗了大量内存。 请告诉我内存泄漏问题在哪里。提前谢谢。 我附上jprofiler快照

Answer 1

找一个调试器并逐步完成循环。我猜其中一个对象没有在每次迭代中清除内存。如果找不到，可能需要一次分块加载，比如一次加载200条记录

通常将数据预取到数组中，并以最快的速度进行处理。在Java中，对象可能非常重