选择在两个表之间连接的查询

Question

我必须以年度格式获取结果，但以下查询返回意外输出：

    SELECT EXTRACT (YEAR FROM Tab_1.DATE_STMP), COUNT (Tab_1.DATE_STMP), 

    EXTRACT (YEAR FROM Tab_2.DATE_STMP), COUNT (Tab_2.DATE_STMP) FROM  

    EMP_1 Tab_1 FULL OUTER JOIN EMP_2@LINK Tab_2  

    ON EXTRACT(YEAR FROM Tab_1.DATE_STMP)= EXTRACT(YEAR FROM Tab_2.DATE_STMP)  

    GROUP BY EXTRACT (YEAR FROM Tab_1.DATE_STMP),

    EXTRACT (YEAR FROM Tab_2.DATE_STMP);

预期输出：

2014    40  2014    20

实际输出：

2014    800 2014    800

我的长 plsql 解决方案是迭代 DATE 值并分别查询 Count：

--PL/SQL
OPEN cur FOR 'SELECT DATE_STMP FROM EMP_1@LINK1 UNION SELECT DATE_STMP FROM EMP_2@LINK2'
FETCH cur INTO v_datVar;
...
EXECUTE IMMEDIATE 'SELECT COUNT(*) FROM EMP_1 WHERE EXTRACT (YEAR FROM DATE_STMP)='||v_datVar INTO v_Total2;
EXECUTE IMMEDIATE 'SELECT COUNT(*) FROM EMP_2 WHERE EXTRACT (YEAR FROM DATE_STMP)='||v_datVar INTO v_Total2;
...
--PRINT VALUES HERE

问题。

是否有单个选择查询来获取此输出？ （使用不同类型的连接或连接条件？）

Answer 1

您在 DATE_STMP 上完成了完全外部联接记录。所以你会得到匹配和不匹配（即日期在另一个表中没有匹配）。您的结果表明您甚至没有匹配项，而只有不匹配项。但是，您似乎对有多少日期匹配完全不感兴趣。此外，我敢说您似乎甚至对年份匹配的频率都不感兴趣。您只想计算这两个表中每年的记录。这与匹配记录和连接无关。

所以：首先获取计数。然后将结果加入，以便每年获得一条记录：

select coalesce(t1.the_year,t2.the_year), t1.the_count, t2.the_count
from
(
  select extract (year from date_stmp) the_year, count(*) as the_count 
  from emp_1
  group by extract (year from date_stmp)
) t1
full outer join
(
  select extract (year from date_stmp) the_year, count(*) as the_count 
  from emp_2@link on
  group by extract (year from date_stmp)
) t2 on t2.the_year = t1.the_year;

编辑：我看到您更改了请求的查询并按年加入。不过，如前所述，我认为您对计算匹配和不匹配不感兴趣，所以我的答案保持不变。

Answer 2

另一种方法是将union all 与聚合一起使用：

SELECT y1, cnt1, y2, cnt2
FROM ((SELECT EXTRACT(YEAR FROM Tab_1.DATE_STMP) as y1, COUNT(*) as cnt1, NULL as y2, NULL as cnt2
       FROM EMP_1 Tab_1
       GROUP BY EXTRACT(YEAR FROM Tab_1.DATE_STMP)
       EMP_2@LINK Tab_2  
      ) UNION ALL
      (SELECT NULL, NULL, EXTRACT(YEAR FROM Tab_2.DATE_STMP) as y2, COUNT(*) as cnt2
       FROM EMP_1 Tab_2
       GROUP BY EXTRACT(YEAR FROM Tab_2.DATE_STMP)
      )
     )  
GROUP BY COALESCE(y1, y2)

Answer 3

我认为最好将每个表分组在单独的select：

SELECT TAB_1.YEAR1, TAB_1.COUNT1, TAB_2.YEAR2, TAB_2.COUNT2

FROM 
(SELECT EXTRACT (YEAR FROM DATE_STMP) AS YEAR1, COUNT(DATE_STMP) AS COUNT1
FROM EMP_1
GROUP BY EXTRACT (YEAR FROM DATE_STMP)) AS TAB_1

FULL OUTER JOIN

(SELECT EXTRACT (YEAR FROM DATE_STMP) AS YEAR2, COUNT(DATE_STMP) AS COUNT2
FROM EMP_2
GROUP BY EXTRACT (YEAR FROM DATE_STMP)) AS TAB_2

ON TAB_1.YEAR1=TAB_2.YEAR2;