在BigQuery中使用“保留”方法（类似于SAS）

Question

在 BigQuery 中定义变量时，如何将一行中的值与上面获得的值进行比较（当变量创建仍在进行中时）？

我有一个包含 ID、开始日期和结束日期的表格。我想创建一个 MaxDate 变量，它将比较（从第二行开始）每行的开始日期是否为 <= to "lag" end date, and if the condition is true, then MaxDate should be the maximum of the EndDate and the "previous" EndDate immediately above.

身份证	开始日期	结束日期	最大日期
A	2019-10-25	2019-10-31	2019-10-31
A	2019-10-26	2019-10-26	2019-10-31
A	2019-10-28	2019-10-30	2019-10-31
A	2019-10-29	2019-10-29	2019-10-31

以下方法在第二个滞后行失败：

WITH
S1 AS
(
SELECT ID, 1 AS COUNT
FROM S0
ORDER BY ID, StartDate, EndDate
)

,S2 AS
(
SELECT *,

CASE WHEN StartDate < LAG(EndDate) OVER (PARTITION BY ID ORDER BY StartDate, EndDate)
THEN
  MAX(EndDate) OVER (PARTITION BY ID ORDER BY StartDate, EndDate
    ROWS BETWEEN 1 PRECEDING AND CURRENT ROW)
      ELSE EndDate
END AS EndDate2
FROM S1
)

,S3 AS
(SELECT *,

CASE WHEN StartDate > LAG(EndDate2) OVER (PARTITION BY ID ORDER BY StartDate, EndDate) + 1
THEN 
  MAX(COUNT) OVER (PARTITION BY ID ORDER BY StartDate, EndDate
    ROWS 1 PRECEDING) + 1
      ELSE 1
END AS Visit_Count

FROM S2
ORDER BY ID, StartDate, EndDate
)

(
SELECT ID, Visit_Count, MIN(StartDate) AS StartDate, MAX(EndDate) AS EndDate
FROM S3
GROUP BY ID, Visit_Count
ORDER BY ID, Visit_Count
)
;

谢谢！

Answer 1

弄清楚：

SELECT *,
CASE WHEN StartDate <= LAG(EndDate) 
  OVER (PARTITION BY PATIENT_ID ORDER BY StartDate) 
      THEN MAX(EndDate)
        OVER (PARTITION BY PATIENT_ID ORDER BY StartDate)
            ELSE StartDate
              END AS MaxDate
FROM table
;