java实现innerjoin关联算法

场景：有三个List<Map>变量，分别为list1、list2、list3

List<Map<String,Object>> list1 = new ArrayList<>();
List<Map<String,Object>> list2 = new ArrayList<>();
List<Map<String,Object>> list3 = new ArrayList<>();

现在需要根据一个输入的字符"list1.column1=list2.column2,list1.column3=list3.column4"（不是固定的）来实现inner join关系的控制，即list1中的map和list2中map通过key值column1和column2关联，同时list1中的map和list3中map通过key值column3和column4关联。在sql中就是"select * from list1 inner join list2 on list1.column1=list2.column2 inner join list3 on list1.column3=list3.column4"，将不同list中的map关联到一行上：

    private static void test(){
        long t0 = System.currentTimeMillis();
        List<Map<String,Object>> list1 = new ArrayList<>();
        List<Map<String,Object>> list2 = new ArrayList<>();
        List<Map<String,Object>> list3 = new ArrayList<>();
        //构造10万条数据
        for (int i = 0; i < 100000; i++) {
            String uuid = UUID.randomUUID().toString();

            Map<String,Object> m1 = new HashMap<>();
            m1.put("id1",i);
            m1.put("value1", uuid.charAt(10)+uuid.charAt(12));

            Map<String,Object> m2 = new HashMap<>();
            m2.put("id2",i);
            m2.put("value2", uuid.charAt(10)+uuid.charAt(1));

            Map<String,Object> m3 = new HashMap<>();
            m3.put("id3",i);
            m3.put("value3", uuid.charAt(10)+uuid.charAt(12));

            list1.add(m1);
            list2.add(m2);
            list3.add(m3);
        }
        long t1 = System.currentTimeMillis();
        System.out.println("构造数据耗时："+(t1-t0));
        //添加关联条件
        ArrayList<JoinRelation> joinRelations = new ArrayList<>();
        joinRelations.add(new JoinRelation("list1.id1=list2.id2"));
        joinRelations.add(new JoinRelation("list1.value1=list2.value2"));
        joinRelations.add(new JoinRelation("list2.id2=list3.id3"));
        joinRelations.add(new JoinRelation("list2.value2=list3.value3"));

        InnerJoiner<Map<String,Object>> joiner = new InnerJoiner<>(joinRelations);
        //字段逻辑
        joiner.setGetter((o, columnName) -> MapUtils.getString(o,columnName,""));

        long t11 = System.currentTimeMillis();
        System.out.println("构造join对象耗时："+(t11-t1));

        //添加数据集
        joiner.addUnit("list1",list1)
                .addUnit("list2",list2)
                .addUnit("list3",list3);
        //进行关联操作
        joiner.join();

        DataKey key = new DataKey("list1");
        //通过key拿到关联结果集，每个结果集数据条数一样
        DataUnit<Map<String, Object>> result = joiner.getResult(key);
        System.out.println("关联条数："+result.size());

        long t2 = System.currentTimeMillis();
        System.out.println("关联耗时："+(t2-t11));
        //遍历其中一个数据集
        for (Map<String, Object> map : result) {
            //找到与该行关联的数据集，如果关联到多条，只会返回第一条
            Map<DataKey, Map<String, Object>> row = joiner.findRow(map, key);
            int i = 0;
        }
        long t3 = System.currentTimeMillis();
        System.out.println("单行遍历查找耗时："+(t3-t2));
    }

测试效果： 

总共10万条数据，通过关联条件得到6千条数据，遍历每一行拿到与之关联的数据总共耗时38秒，平均每次关联耗时6毫秒，后面有空会进行优化

 源码链接：https://download.csdn.net/download/qq_36635569/87770897