使用时间码/时间范围聚合路径元素

问题描述 投票：0回答：1

我有以下代码。最后我想总结一下我的原始数据。应汇总重叠周期并输出最小/最大日期。

下面的代码给出了每个ID的最长PATH。到目前为止，一切都很好。不幸的是，我不知道如何使用 ID 来根据组指定最长的整体 PATH。

原表：

身份证 达腾 BEGINN_ZEITSTEMPEL ENDE_ZEITSTEMPEL 团体
1 数据1A 2023-05-01 00:00:00 2023-06-01 00:00:00 1
2 数据2A 2023-04-01 00:00:00 2023-05-15 00:00:00 1
3 数据3A 2023-03-01 00:00:00 2023-05-30 00:00:00 1
4 数据4A 2023-02-01 00:00:00 2023-05-29 00:00:00 1
5 数据5A 2023-05-14 00:00:00 2023-05-15 00:00:00 2
6 数据1B 2023-05-29 00:00:00 2023-05-30 00:00:00 2
7 数据2B 2023-05-01 00:00:00 2023-08-01 00:00:00 3
8 数据3B 2023-05-01 00:00:00 2023-09-01 00:00:00 3
9 数据4B 2023-05-01 00:00:00 2023-06-01 00:00:00 3
10 数据5B 2021-05-01 00:00:00 2022-06-01 00:00:00 3

决赛桌：

START_ID 达腾 BEGINN_ZEITSTEMPEL ENDE_ZEITSTEMPEL 团体 路径
1 数据1A 2023-02-01 00:00:00 2023-06-01 00:00:00 1 2 -> 4 -> 3 -> 1
5 数据5A 2023-05-14 00:00:00 2023-05-15 00:00:00 2 5
6 数据1B 2023-05-29 00:00:00 2023-05-30 00:00:00 2 6
7 数据3B 2023-05-01 00:00:00 2023-09-01 00:00:00 3 9 -> 7 -> 8
10 数据5B 2021-05-01 00:00:00 2022-06-01 00:00:00 3 10 
    WITH CTE as (
            Select 1 as ID, 'Datensatz1A' as Daten, TO_DATE('2023-05-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-06-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 1 as Gruppe FROM DUAL
            UNION Select 2 as ID, 'Datensatz2A' as Daten, TO_DATE('2023-04-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-05-15 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 1 as Gruppe FROM DUAL
            UNION Select 3 as ID, 'Datensatz3A' as Daten, TO_DATE('2023-03-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-05-30 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 1 as Gruppe FROM DUAL
            UNION Select 4 as ID, 'Datensatz4A' as Daten, TO_DATE('2023-02-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-05-29 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 1 as Gruppe FROM DUAL
            UNION Select 5 as ID, 'Datensatz5A' as Daten, TO_DATE('2023-05-14 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-05-15 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 2 as Gruppe FROM DUAL
            UNION Select 6 as ID, 'Datensatz1B' as Daten, TO_DATE('2023-05-29 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-05-30 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 2 as Gruppe FROM DUAL
            UNION Select 7 as ID, 'Datensatz2B' as Daten, TO_DATE('2023-05-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-08-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 3 as Gruppe FROM DUAL
            UNION Select 8 as ID, 'Datensatz3B' as Daten, TO_DATE('2023-05-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-09-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 3 as Gruppe FROM DUAL
            UNION Select 9 as ID, 'Datensatz4B' as Daten, TO_DATE('2023-05-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2023-06-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 3 as Gruppe FROM DUAL
            UNION Select 10 as ID, 'Datensatz5B' as Daten, TO_DATE('2021-05-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Beginn_Zeitstempel, TO_DATE('2022-06-01 00:00:00', 'YYYY-MM-DD HH24:MI:SS') as Ende_Zeitstempel, 3 as Gruppe FROM DUAL
    ),
    RecursiveCTE (Start_ID, End_ID, Daten, Beginn_Zeitstempel, Ende_Zeitstempel, Gruppe, Path) AS (
        SELECT
            ID,
            ID,
            Daten,
            Beginn_Zeitstempel,
            Ende_Zeitstempel,
            Gruppe,
            CAST(ID AS VARCHAR2(100)) AS PATH
        FROM CTE
        WHERE NOT EXISTS (
            SELECT 1
            FROM CTE c
            WHERE c.ID = CTE.ID
            AND c.Gruppe = CTE.Gruppe
            AND c.Beginn_Zeitstempel < CTE.Beginn_Zeitstempel
        )
        UNION ALL
        SELECT
            r.Start_ID,
            c.ID,
            c.Daten,
            LEAST(r.Beginn_Zeitstempel, c.Beginn_Zeitstempel),
            GREATEST(r.Ende_Zeitstempel, c.Ende_Zeitstempel),
            c.Gruppe,
            r.Path || ' -> ' || c.ID AS Pfad
            --r.Path || c.ID AS PATH        
        FROM RecursiveCTE r
        JOIN CTE c ON r.End_ID <> c.ID
                    AND r.Gruppe = c.Gruppe
                    AND r.Ende_Zeitstempel >= c.Beginn_Zeitstempel
                    AND r.Ende_Zeitstempel <= c.Ende_Zeitstempel
                    AND INSTR(r.Path, c.ID) = 0
    ),
    -- SELECT    * from RecursiveCTE
    MaxPathCTE AS (
        SELECT
            Start_ID,
            Gruppe,
            MAX(PATH) KEEP (DENSE_RANK LAST ORDER BY LENGTH(PATH)) AS MAX_PATH
        FROM RecursiveCTE
        GROUP BY Start_ID, Gruppe
    )
    SELECT
        r.Start_ID,
        r.Daten,
        r.Beginn_Zeitstempel,
        r.Ende_Zeitstempel,
        r.Gruppe,
        r.Path
    FROM RecursiveCTE r
    INNER JOIN MaxPathCTE m ON r.Start_ID = m.Start_ID AND r.Gruppe = m.Gruppe AND r.Path = m.MAX_PATH
    ORDER BY r.Start_ID;
sql oracle common-table-expression
1个回答
0
投票

从 Oracle 12 开始，您可以使用 

MATCH_RECOGNIZE
进行逐行模式匹配：

SELECT *
FROM   table_name
MATCH_RECOGNIZE (
  PARTITION BY Gruppe
  ORDER     BY Ende_Zeitstempel DESC, Beginn_Zeitstempel DESC
  MEASURES
    FIRST(id) AS start_id,
    LAST(ID) AS last_id,
    COUNT(*) AS num_rows,
    MIN(Beginn_Zeitstempel) AS Beginn_Zeitstempel,
    FIRST(Ende_Zeitstempel) AS Ende_Zeitstempel
  PATTERN (overlapping* non_overlapping)
  DEFINE
    overlapping AS MIN(Beginn_Zeitstempel) <= NEXT(Ende_Zeitstempel)
)

对于样本数据：

CREATE TABLE table_name (id, daten, Beginn_Zeitstempel, Ende_Zeitstempel, Gruppe) AS
  Select  1, 'Datensatz1A', DATE '2023-05-01', DATE '2023-06-01', 1 FROM DUAL UNION ALL
  Select  2, 'Datensatz2A', DATE '2023-04-01', DATE '2023-05-15', 1 FROM DUAL UNION ALL
  Select  3, 'Datensatz3A', DATE '2023-03-01', DATE '2023-05-30', 1 FROM DUAL UNION ALL
  Select  4, 'Datensatz4A', DATE '2023-02-01', DATE '2023-05-29', 1 FROM DUAL UNION ALL
  Select  5, 'Datensatz5A', DATE '2023-05-14', DATE '2023-05-15', 2 FROM DUAL UNION ALL
  Select  6, 'Datensatz1B', DATE '2023-05-29', DATE '2023-05-30', 2 FROM DUAL UNION ALL
  Select  7, 'Datensatz2B', DATE '2023-05-01', DATE '2023-08-01', 3 FROM DUAL UNION ALL
  Select  8, 'Datensatz3B', DATE '2023-05-01', DATE '2023-09-01', 3 FROM DUAL UNION ALL
  Select  9, 'Datensatz4B', DATE '2023-05-01', DATE '2023-06-01', 3 FROM DUAL UNION ALL
  Select 10, 'Datensatz5B', DATE '2021-05-01', DATE '2022-06-01', 3 FROM DUAL

输出：

团体 START_ID LAST_ID NUM_ROWS BEGINN_ZEITSTEMPEL ENDE_ZEITSTEMPEL
1 1 2 4 2023-02-01 00:00:00 2023-06-01 00:00:00
2 6 6 1 2023-05-29 00:00:00 2023-05-30 00:00:00
2 5 5 1 2023-05-14 00:00:00 2023-05-15 00:00:00
3 8 9 3 2023-05-01 00:00:00 2023-09-01 00:00:00
3 10 10 1 2021-05-01 00:00:00 2022-06-01 00:00:00

小提琴

最新问题
© www.soinside.com 2019 - 2024. All rights reserved.