Question

问题

需要填补数据空白。如果可能，请避免UNION或PARTITION BY。

查询声明

select语句如下：

SELECT
  count( r.incident_id ) AS incident_tally,
  r.severity_cd,
  r.incident_typ_cd
FROM
  report_vw r
GROUP BY
  r.severity_cd, r.incident_typ_cd
ORDER BY
  r.severity_cd,
  r.incident_typ_cd

数据来源

严重性代码和事件类型代码来自：

severity_vw
incident_type_vw

列是：

incident_tally
severity_cd
incident_typ_cd

实际结果数据

36  0   ENVIRONMENT
1   1   DISASTER
27  1   ENVIRONMENT
4   2   SAFETY
1   3   SAFETY

必填结果数据

36  0   ENVIRONMENT
0   0   DISASTER
0   0   SAFETY
27  1   ENVIRONMENT
0   1   DISASTER
0   1   SAFETY
0   2   ENVIRONMENT
0   2   DISASTER
4   2   SAFETY
0   3   ENVIRONMENT
0   3   DISASTER
1   3   SAFETY

问题

您如何使用UNION，PARTITION BY或LEFT JOIN填写零点数？

Answer 1

像

这样的东西

SELECT  COUNT(r.incident_id),
        crsjn.severity_cd,
        crsjn.incident_typ_cd
FROM    (
            SELECT  severity_cd,
                    incident_typ_cd
            FROM    severity_vw,
                    incident_type_vw
        ) crsjn LEFT JOIN
        report_vw r     ON  crsjn.severity_cd = r.severity_cd 
                        AND crsjn.incident_typ_cd = r.incident_typ_cd
GROUP BY crsjn.severity_cd,
        crsjn.incident_typ_cd
ORDER BY crsjn.severity_cd,
        crsjn.incident_typ_cd

Answer 2

最容易的是，如果您在另一个表中有事件类型和事件严重性。

SELECT COALESCE(sub.incident_tally), s.severity_cd, i.incident_type_cd
FROM incident_type i
CROSS JOIN incident_severity s
LEFT JOIN (
   ... your original sql statement...
) sub ON i.incident_typ_cd = sub.incident_type_cd
  AND s.severity_cd = sub.severity_cd

如果你没有这些表，你可以做这样的事情。我不明白为什么没有UNION限制。但这很有效。

CREATE TYPE VARCHAR_TABLE AS TABLE OF VARCHAR2(60);
CREATE TYPE NUMBER_TABLE AS TABLE OF NUMBER;

WITH inc AS (
  SELECT /*+ CARDINALITY(nt1 3) */ column_value AS incident_type_cd
  FROM TABLE(varchar_table('ENVIRONMENT', 'DISASTER', 'SAFETY')) nt1
),
sev AS (
  SELECT /*+ CARDINALITY(nt2 4) */ column_value AS severity_cd
  FROM TABLE(number_table(0,1,2,3)) nt2
) 
SELECT *
FROM inc
CROSS JOIN sev
LEFT JOIN (
   ... your original sql statement...
) sub ON i.incident_typ_cd = sub.incident_type_cd
  AND s.severity_cd = sub.severity_cd

4

Answer 3

我能想到的唯一方法是进一步规范化该表并使用左连接。

填补数据空白 - UNION，PARTITION BY或JOIN？

3 个答案: