Oracle SQL - 将表行数据透视表并在数据透视表中使用子查询

Question

我正在开发 Oracle 12c R1 db，并有一个包含示例数据的示例视图，如下所示：
视图名称：CUST_HOTEL_VIEW

+----------------+---------------+---------------+
|    Customer    |     Hotel     | Booked Status |
+----------------+---------------+---------------+
| John Smith     | Beverly Hills | Booked        |
| John Smith     | Royal Palms   |               |
| Marilyn Lawson | Beverly Hills |               |
| John Smith     | Ritz-Carlton  |               |
| Marilyn Lawson | Royal Palms   |               |
| Sarah Elliot   | Royal Palms   |               |
| Sarah Elliot   | Ritz-Carlton  | Booked        |
| Sarah Elliot   | Royal Palms   | Booked        |
+----------------+---------------+---------------+

根据上面的数据，我试图通过行总计、列总计和每位客户预订的酒店数量低于枢轴输出：

+----------------+-------------+---------------+--------------+-------------+----------+
|    Customer    | Royal Palms | Beverly Hills | Ritz-Carlton | Grand Total | # Booked |
+----------------+-------------+---------------+--------------+-------------+----------+
| John Smith     |           1 |             1 |            1 |           3 |        1 |
| Marilyn Lawson |           1 |             1 |              |           2 |        - |
| Sarah Elliot   |           2 |               |            1 |           3 |        2 |
| Grand Total    |           4 |             2 |            2 |           8 |        3 |
+----------------+-------------+---------------+--------------+-------------+----------+

我尝试了下面的查询来生成数据透视数据

SELECT * FROM
(
  SELECT CUSTOMER, HOTEL
  FROM CUST_HOTEL_VIEW
)
PIVOT
(
  COUNT(HOTEL)
  FOR HOTEL IN ('Royal Palms' as "Royal Palms",'Beverly Hills' as "Beverly Hills",'Ritz-Carlton' as "Ritz-Carlton")
)
ORDER BY CUSTOMER

我想知道：
1. 如何包含行总计
2. 如何包含列总计
3.如何包括预订酒店数量和
3. 是否可以在 PIVOT FOR HOTEL IN 子句中编写子查询。 （我尝试了子查询但出现错误）

我很感激这方面的任何帮助。

谢谢，
里查

Answer 1

只使用条件聚合：

SELECT COALESCE(customer, 'Grand Total') as customer,
       SUM(CASE WHEN Hotel = 'Royal Palms' THEN 1 ELSE 0 END) as "Royal Palms",
       SUM(CASE WHEN Hotel = 'Beverly Hills' THEN 1 ELSE 0 END) as "Beverly Hills",       
       SUM(CASE WHEN Hotel = 'Ritz-Carlton' THEN 1 ELSE 0 END) as "Ritz-Carlton" ,
       COUNT(*) as "Grand Total",
       COUNT(Booked_Status) as "Num Booked"
FROM CUST_HOTEL_VIEW
GROUP BY ROLLUP(CUSTOMER)
ORDER BY CUSTOMER;

条件聚合比pivot 灵活得多。就个人而言，我认为pivot 语法没有任何理由：它做得很好，但不像传统 SQL 语句那样是构建块。

ROLLUP() 也很有帮助。您还可以使用：

GROUP BY GROUPING SETS ( (CUSTOMER), () )

Answer 2

可能有一个更简单的解决方案，但这应该可以帮助您入门 -

WITH A AS (
    SELECT
        *
    FROM
        (
            SELECT
                CUSTOMER,
                HOTEL,
                BOOKED_STATUS
            FROM
                TABLE1
        )
            PIVOT ( COUNT ( HOTEL )
                FOR HOTEL
                IN ( 'ROYAL PALMS' AS "ROYAL PALMS",'BEVERLY HILLS' AS "BEVERLY HILLS",'RITZ-CARLTON' AS "RITZ-CARLTON" )
            )
),B AS (
    SELECT
        CUSTOMER,
        "ROYAL PALMS",
        "BEVERLY HILLS",
        "RITZ-CARLTON",
        SUM("ROYAL PALMS" + "BEVERLY HILLS" + "RITZ-CARLTON") AS "GRAND TOTAL"
    FROM
        A
    GROUP BY
        CUSTOMER,
        "ROYAL PALMS",
        "BEVERLY HILLS",
        "RITZ-CARLTON"
),C AS (
    SELECT
        CUSTOMER,
        SUM("ROYAL PALMS" + "BEVERLY HILLS" + "RITZ-CARLTON") AS NBOOK
    FROM
        A
    WHERE
        BOOKED_STATUS IS NOT NULL
    GROUP BY
        CUSTOMER
),D AS (
    SELECT
        B.CUSTOMER,
        SUM("ROYAL PALMS") AS "ROYAL PALMS",
        SUM("BEVERLY HILLS") AS "BEVERLY HILLS",
        SUM("RITZ-CARLTON") AS "RITZ-CARLTON",
        SUM("GRAND TOTAL") AS "GRAND TOTAL",
        NVL(C.NBOOK,0) AS NBOOK
    FROM
        B,
        C
    WHERE
        B.CUSTOMER   = C.CUSTOMER (+)
    GROUP BY
        B.CUSTOMER,
        NVL(C.NBOOK,0)
) SELECT
    *
FROM
    D
UNION
SELECT
    'GRAND TOTAL' AS CUSTOMER,
    SUM("ROYAL PALMS"),
    SUM("BEVERLY HILLS"),
    SUM("RITZ-CARLTON"),
    SUM("GRAND TOTAL") AS "GRAND TOTAL",
    SUM(NBOOK) AS NBOOK
FROM
    D
ORDER BY 5,1;

输出 -

"CUSTOMER","ROYAL PALMS","BEVERLY HILLS","RITZ-CARLTON","GRAND TOTAL","NBOOK"
"MARILYN LAWSON",1,1,0,2,0
"JOHN SMITH",1,1,1,3,1
"SARAH ELLIOT",2,0,1,3,2
"GRAND TOTAL",4,2,2,8,3

Answer 3

试试这个：

1) 包含行总数union 可以使用

2) 在第一个表 (p1) 中使用 sum 函数来包含列总计：与 sum 函数编写的代码相同

3) 要包含预订酒店的数量，another pivot 是必填项，即第二张表 (p2)

---

Step1：制作一张表 p1，包含客户、不同酒店及其总计（列总计）

---

create table p1 as
SELECT customer,RoyalPalms,BeverlyHills,RitzCarlton,sum(RoyalPalms + BeverlyHills + RitzCarlton) as GrandTotal  FROM
(
  SELECT CUSTOMER, HOTEL
  FROM CUST
)
PIVOT
( COUNT(HOTEL)
  FOR HOTEL IN ('Royal Palms' as RoyalPalms,'Beverly Hills' as BeverlyHills,
                'Ritz-Carlton' as RitzCarlton) ) 
group by customer,RoyalPalms,BeverlyHills,RitzCarlton
order by customer;                

---

第 2 步：与客户和预订的酒店制作一张表 p2

---

create table p2 as
SELECT *  FROM
(
  SELECT customer,booked_status
  FROM cust
)
pivot 
 ( count(booked_status) 
  for booked_status in ('Booked' as Booked));

---

Step3：连接表p1和p2

Step4：对于行合计，取Step3中的表与汇总函数计算的行合计的并集

---

(SELECT a.*,b.booked
 from p1 a 
 left join
 p2 b
 on a.customer = b.customer) 
union all
SELECT 'GrandTotal', Sum(RoyalPalms),sum(BeverlyHills),sum(RitzCarlton),
sum(GrandTotal),sum(booked) From test;

输出：

+----------------+-------------+---------------+--------------+-------------+----------+
|    Customer    | Royal Palms | Beverly Hills | Ritz-Carlton | Grand Total | # Booked |
+----------------+-------------+---------------+--------------+-------------+----------+
| John Smith     |           1 |             1 |            1 |           3 |        1 |
| Marilyn Lawson |           1 |             1 |              |           2 |        - |
| Sarah Elliot   |           2 |               |            1 |           3 |        2 |
| Grand Total    |           4 |             2 |            2 |           8 |        3 |
+----------------+-------------+---------------+--------------+-------------+----------+

如果需要任何进一步的查询/解释，请告诉我。