计数值逗号分隔

Question

我有下表：

A   V
-   -
A1  B, C
A2  C, 2W
A3  D
A4  10K, 10W, 2A

我想计算 V 中的值并按出现次数排序。预期输出：

O  COUNT(*)
-  -------
W       12
K       10
A        2
C        2
B        1
D        1

Answer 1

您可以将 REGEX_SUBSTR() 和 REGEXP_REPLACE() 函数与分层查询结合使用，假设所有以逗号分隔的元素，无论何时修剪空格，都是 正整数形式+ 一个字母，例如

WITH t2 AS
(
 SELECT TRIM(REGEXP_SUBSTR(v,'[^,]+',1,level)) AS v
   FROM t 
CONNECT BY level <= REGEXP_COUNT(v,',')+1
    AND PRIOR SYS_GUID() IS NOT NULL
    AND PRIOR a = a
)
SELECT NVL(SUM(REGEXP_REPLACE(v,'[^[:digit:]]')),
           CASE WHEN COUNT(*)>1 THEN COUNT(*) END)||REGEXP_REPLACE(v,'[^[:alpha:]]') AS o
  FROM t2 
 GROUP BY REGEXP_REPLACE(v,'[^[:alpha:]]')    

Demo

编辑（取决于问题的最后编辑）：只需删除连接运算符，并为 COUNT 聚合设置条件，例如

SELECT REGEXP_REPLACE(v,'[^[:alpha:]]') AS o,
       NVL(SUM(REGEXP_REPLACE(v,'[^[:digit:]]')),COUNT(*)) AS count 
  FROM t2 
 GROUP BY REGEXP_REPLACE(v,'[^[:alpha:]]') 
 ORDER BY count DESC, o

Demo