POI 解析百万行 excel 的实现

poi 的usermodel api读取大数据量excel会导致OOM，可以使用eventmodel api来处理这种excel.


少量的行数excel可以用

XSSFWorkbook wb = new XSSFWorkbook(inputStream);  
         XSSFSheet sheet = wb.getSheetAt(0);  
         Iterator<Row> iter = sheet.iterator();  
         boolean isfirstline = true;  
         while (iter.hasNext()) {  
             Row row = iter.next();  
             if (isfirstline) { // 忽略上传文件第一行的标题栏  
                 isfirstline = false;  
                 continue;  
             }  
              //解析excel，每行有11列，然后对每列解析出来之后调用后端服务把数据保存到数据库中，  
             }  
         } 

XSSFWorkbook wb = new XSSFWorkbook(inputStream);  

new ExcelEventUserModelParse(filePath)
                    .setHandler(new ExcelEventUserModelParse.SimpleSheetContentsHandler() {
                        @Override
                        public void endRow(int rowNum) {
                            // System.out.println(row.toString());
                            if (!isEmpty.getIsEmpty()) {
                                table.add(row);
                            } else {
                                isnull.setIsEmpty(isEmpty.getIsEmpty());
                                isnull.setRowNum(isEmpty.getRowNum());
                            }
                        }
                    }).parse();

public class ExcelEventUserModelParse {
    private Logger logger = LoggerFactory.getLogger(ExcelEventUserModelParse.class);
    private String filename;
    private SheetContentsHandler handler;
public ExcelEventUserModelParse(String filename) {
    this.filename = filename;
}


public ExcelEventUserModelParse setHandler(SheetContentsHandler handler) {
    this.handler = handler;
    return this;
}


public void parse() {
    OPCPackage pkg = null;
    InputStream sheetInputStream = null;
    try {
        pkg = OPCPackage.open(filename, PackageAccess.READ);
        XSSFReader xssfReader = new XSSFReader(pkg);
        StylesTable styles = xssfReader.getStylesTable();
        ReadOnlySharedStringsTable strings = new ReadOnlySharedStringsTable(pkg);
        sheetInputStream = xssfReader.getSheetsData().next();
        processSheet(styles, strings, sheetInputStream);
    } catch (Exception e) {
        logger.error(e.getMessage());
        throw new RuntimeException(e.getMessage(), e);
    } finally {
        if (sheetInputStream != null) {
            try {
                sheetInputStream.close();
            } catch (IOException e) {
                logger.error(e.getMessage());
                throw new RuntimeException(e.getMessage(), e);
            }
        }
        if (pkg != null) {
            try {
                pkg.close();
            } catch (IOException e) {
                logger.error(e.getMessage());
                throw new RuntimeException(e.getMessage(), e);
            }
        }
    }
}


private void processSheet(StylesTable styles, ReadOnlySharedStringsTable strings, InputStream sheetInputStream)
        throws SAXException, ParserConfigurationException, IOException {
    XMLReader sheetParser = SAXHelper.newXMLReader();


    if (handler != null) {
        sheetParser.setContentHandler(new XSSFSheetXMLHandler(styles, strings, handler, false));
    } else {
        sheetParser.setContentHandler(
                new XSSFSheetXMLHandler(styles, strings, new SimpleSheetContentsHandler(), false));
    }


    sheetParser.parse(new InputSource(sheetInputStream));
}


public static class SimpleSheetContentsHandler implements SheetContentsHandler {
    private Logger logger = LoggerFactory.getLogger(SimpleSheetContentsHandler.class);
    protected ImportDecryptModel row = new ImportDecryptModel();
    protected int a = -1;
    protected JudgementModel isEmpty = new JudgementModel();


    @Override
    public void startRow(int rowNum) {
        if (rowNum == a + 1) {
            a = rowNum;
        } else {
            isEmpty.setIsEmpty(true);
            logger.error("第" + rowNum + "行数据为空");
            a = rowNum;
            isEmpty.setRowNum(rowNum);
        }
        row = new ImportDecryptModel();
    }


    @Override
    public void endRow(int rowNum) {
        logger.error(rowNum + " : " + row);
    }


    @Override
    public void cell(String cellReference, String formattedValue, XSSFComment comment) {
        if (isEmpty != null && !isEmpty.getIsEmpty()) {
            switch (cellReference.substring(0, 1)) {
            case "A":
                row.setUserId(formattedValue);
                break;
            case "B":
                row.setUid(formattedValue);
                break;
            case "C":
                row.setUserName(formattedValue);
                break;
            case "D":
                row.setIdCard(formattedValue);
                break;
            case "E":
                row.setMobileNo(formattedValue);
                break;
            case "F":
                row.setBankCard(formattedValue);
                break;
            default:
                break;
            }
        }
    }


    @Override
    public void headerFooter(String text, boolean isHeader, String tagName) {


    }
}</pre>



解析的工作全部在cell（）里面，速度大约100万条记录20s左右。关键不会内存溢出