HBase根据Rowkey批量查询数据JAVA API

最近在生产中遇到了一个需求，前台给我多个rowkey的List，要在hbase中查询多个记录（返回给前台list）。在网上也查了很多，不过自己都不太满意，filter的功能有可能查询结果不是准确值，而网上给出的get方法也都是返回一条，scan的话都是返回全部数据，还有用rowkey范围查询的，都跟我的这个应用场景不符啊。无奈，自己找了一个方法，给各位有同样需求的朋友们一个参考。

首先创建链接属性：

    public static Configuration conf = null;
        public static Connection connection = null;
        public static Admin admin = null;
        static {
            conf = HBaseConfiguration.create();
     
            conf.set("hbase.zookeeper.quorum", "10.120.193.800,10.120.193.810");
            conf.set("hbase.master", "10.120.193.730:60010");
            conf.set("hbase.zookeeper.property.clientPort", "2181");
            conf.set("zookeeper.znode.parent", "/hbase-unsecure");
            conf.set("hbase.client.keyvalue.maxsize", "1048576000"); //1G
            conf = HBaseConfiguration.create(conf);
            try {
                connection = ConnectionFactory.createConnection(conf);
                admin = connection.getAdmin();
            } catch (IOException e) {
                e.printStackTrace();
            }
        }

紧接着，开始做查询操作:

        public static List<String> qurryTableTest(List<String> rowkeyList) throws IOException {
            String tableName = "table_a";
            Table table = connection.getTable( TableName.valueOf(tableName));// 获取表
            for (String rowkey : rowkeyList){
                Get get = new Get(Bytes.toBytes(rowkey));
                Result result = table.get(get);
                for (Cell kv : result.rawCells()) {
                    String value = Bytes.toString(CellUtil.cloneValue(kv));
                    list.add(value);
                }
            }
            return list;
        }

但是！！！重点来了，每条rowkey都要发起一次请求，这种方法效率十分低，测试了10000条记录查询要用4分钟左右，项目需求是秒回啊，这怎么能行？

于是，就自己点开hbase java 源码，自己慢慢找，忽然眼前一亮！在HTable.class里看到了这个get方法：

    public Result[] get(List<Get> gets) throws IOException {
            if(gets.size() == 1) {
                return new Result[]{this.get((Get)gets.get(0))};
            } else {
                try {
                    Object[] r1 = this.batch(gets);
                    Result[] results = new Result[r1.length];
                    int i = 0;
                    Object[] arr$ = r1;
                    int len$ = r1.length;
     
                    for(int i$ = 0; i$ < len$; ++i$) {
                        Object o = arr$[i$];
                        results[i++] = (Result)o;
                    }
     
                    return results;
                } catch (InterruptedException var9) {
                    throw (InterruptedIOException)(new InterruptedIOException()).initCause(var9);
                }
            }
        }

这就简单了，紧接着对自己上边的方法进行改进：

    public static List<String> qurryTableTestBatch(List<String> rowkeyList) throws IOException {
            List<Get> getList = new ArrayList();
            String tableName = "table_a";
            Table table = connection.getTable( TableName.valueOf(tableName));// 获取表
            for (String rowkey : rowkeyList){//把rowkey加到get里，再把get装到list中
                Get get = new Get(Bytes.toBytes(rowkey));
                getList.add(get);
            }
            Result[] results = table.get(getList);//重点在这，直接查getList<Get>
            for (Result result : results){//对返回的结果集进行操作
                for (Cell kv : result.rawCells()) {
                    String value = Bytes.toString(CellUtil.cloneValue(kv));
                    list.add(value);
                }
            }
            return list;
        }

根据这种批量的方法，10000个rowkey进行测试，时间为2s，速度提升特别大。

ok!大功告成，成功解决了问题~

---------------------
作者：0箫逝
来源：CSDN
原文：https://blog.csdn.net/a867565870/article/details/78683839
版权声明：本文为博主原创文章，转载请附上博文链接！