Sql 在SELECT的结果中包含SELECT未恢复任何数据的行

我正在尝试编写查询以检索以下数据：

 TYPE     |    TOTAL   |  0_10 DAYS   |  10_20 DAYS   |  .......
  X             300         100            200           .......
  Y              0           0              0            .......
  Z             600         50             120           ....... 

我必须按类型对我的所有条目进行分组，并计算每个日期范围内每种类型的条目数量，并将它们合计起来。 我的问题是需要为我没有检索任何数据的类型显示一行零。基本上，“类型”列始终显示固定数量的类型。到目前为止，我已经尝试过使用“UNION ALL”，但是零行总是会显示出来。我的问题是：

SELECT TYPE             AS "ORDERS", 
       Count(*)         AS "TOTAL", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 10 AND SYSDATE ) THEN 1 
                 ELSE 0 
               END), 0) AS "0_10_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 20 AND SYSDATE - 11 ) THEN 
                 1 
                 ELSE 0 
               END), 0) AS "10_20_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 30 AND SYSDATE - 21 ) THEN 
                 1 
                 ELSE 0 
               END), 0) AS "20_30_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER <= To_date(SYSDATE - 30) ) THEN 1 
                 ELSE 0 
               END), 0) AS "PLUS_30_DAYS" 
FROM   T_ORDERS 
WHERE  TYPE = 'X' 
        OR TYPE = 'Y' 
        OR TYPE = 'Z' 
GROUP  BY TYPE 
UNION ALL 
SELECT TYPE AS "ORDERS", 
       0    AS "TOTAL", 
       0    AS "0_10_DAYS", 
       0    AS "10_20_DAYS", 
       0    AS "20_30_DAYS", 
       0    AS "PLUS_30_DAYS" 
FROM   T_ORDERS 
WHERE  TYPE IS NOT NULL 
GROUP  BY TYPE; 

我是SQL新手，所以如果这个话题上的任何问题的答案都能解决我的问题，请容忍我，但我似乎无法解决它。如果有不清楚的地方，请写在评论框中

试试看：

SELECT TYPE              AS "ORDERS", 
       Count(DATE_ORDER) AS "TOTAL", 
       Nvl(Sum(CASE 
                 WHEN (DATE_ORDER BETWEEN SYSDATE - 10 AND SYSDATE) THEN 1 
                 ELSE 0 
               END), 0)  AS "0_10_DAYS", 
       Nvl(Sum(CASE 
                 WHEN (DATE_ORDER BETWEEN SYSDATE - 20 AND SYSDATE - 11) THEN 1
                 ELSE 0 
               END), 0)  AS "10_20_DAYS", 
       Nvl(Sum(CASE 
                 WHEN (DATE_ORDER BETWEEN SYSDATE - 30 AND SYSDATE - 21) THEN 1
                 ELSE 0 
               END), 0)  AS "20_30_DAYS", 
       Nvl(Sum(CASE 
                 WHEN (DATE_ORDER <= To_date(SYSDATE - 30)) THEN 1 
                 ELSE 0 
               END), 0)  AS "PLUS_30_DAYS" 
FROM   (SELECT TYPE, DATE_ORDER 
        FROM T_ORDERS 
        WHERE TYPE IN ('X', 'Y', 'Z')
        UNION ALL
        SELECT DECODE(LEVEL, 1,'X', 2,'Y', 3,'Z') TYPE, NULL DATE_ORDER 
        FROM DUAL
        CONNECT BY LEVEL <= 3 
       ) SQ
GROUP  BY TYPE 

---

可以使用外部联接来获得所需的结果

with types
AS
(
  SELECT 'X' as t_name FROM dual
  UNION ALL
  SELECT 'Y' as t_name FROM dual
  UNION ALL
  SELECT 'Z' as t_name FROM dual
)
SELECT 
       types.t_name     AS "ORDERS", 
       SUM(CASE WHEN T_ORDERS.TYPE IS NULL THEN 0 ELSE 1 END )   AS "TOTAL", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 10 AND SYSDATE ) THEN 1 
                 ELSE 0 
               END), 0) AS "0_10_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 20 AND SYSDATE - 11 ) THEN 
                 1 
                 ELSE 0 
               END), 0) AS "10_20_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 30 AND SYSDATE - 21 ) THEN 
                 1 
                 ELSE 0 
               END), 0) AS "20_30_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER <= To_date(SYSDATE - 30) ) THEN 1 
                 ELSE 0 
               END), 0) AS "PLUS_30_DAYS" 
FROM   
  types 
  LEFT OUTER JOIN T_ORDERS ON (types.t_name = T_ORDERS.TYPE ) 
GROUP  BY types.t_name 

或者，如果类型表中已包含所有类型，则可以使用此表而不是

SELECT 
       types.t_name     AS "ORDERS", 
       SUM(CASE WHEN T_ORDERS.TYPE IS NULL THEN 0 ELSE 1 END )   AS "TOTAL",  
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 10 AND SYSDATE ) THEN 1 
                 ELSE 0 
               END), 0) AS "0_10_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 20 AND SYSDATE - 11 ) THEN 
                 1 
                 ELSE 0 
               END), 0) AS "10_20_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER BETWEEN SYSDATE - 30 AND SYSDATE - 21 ) THEN 
                 1 
                 ELSE 0 
               END), 0) AS "20_30_DAYS", 
       Nvl(Sum(CASE 
                 WHEN ( DATE_ORDER <= To_date(SYSDATE - 30) ) THEN 1 
                 ELSE 0 
               END), 0) AS "PLUS_30_DAYS" 
FROM   
  types 
  LEFT OUTER JOIN T_ORDERS ON (types.t_name = T_ORDERS.TYPE ) 
WHERE 
  types.t_name IN ( 'X', 'Y','Z')
GROUP  BY types.t_name 

---

你能提供一些样本数据吗？你的表格设计是错误的。而不是类型|总计| 0|U 10天| 10|U 20天|。。。您应该输入| DAYS_NUM | VAL。然后问题会自行解决。此解决方案不会为没有数据可检索的每种类型显示一行零。@user2629578:请立即尝试。若你们有一个单独的类型值查找表，你们可以用它来代替选择解码…太棒了！我认为它起作用了。我要再测试一下。你能给我简单解释一下你做了什么吗？我对解码和电平有点迷茫。非常感谢@user2629578:如果运行select decode，可能会更清晰一些。。。查询本身-如此处所示：。decode函数是Oracle案例构造的缩写版本-此处的文档：。connect by子句用于生成从1到3的数字列表；这是一个Oracle构造，也可用于生成分层查询-此处的文档：。类型不在类型表中。我已经尝试了您的第一个解决方案，当没有检索到该类型的数据时，它总共显示1。@user2629578:好的，我已经编辑了我的答案。我已经修改了TOTAL列的表达式