计算时间戳的差异并跳过 NULL

Question

如何跳过第一行来计算差异的平均值？ “WHERE”筛选器不起作用 （错误：WHERE 子句中不允许解析函数）

为什么我可以跳过“557186”的第一个值？
是否可以声明“Unit”以便稍后将此变量用于“WHERE”？

SELECT
  FORMAT_TIMESTAMP("%F", Timestamp) AS Date,
  FORMAT_TIMESTAMP("%R", Timestamp) AS Unit,
  LAG(FORMAT_TIMESTAMP("%R", Timestamp)) OVER (ORDER BY FORMAT_TIMESTAMP("%R", Timestamp)) AS CycleTimePreValue,
  CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64)-LAG(CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64)) OVER (ORDER BY CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64)) AS CycleTimePreValueSeconds,
FROM
  `wh-lr-sk.Quality.Retouren_IMPORT`
  # WHERE (CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64)-LAG(CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64)) OVER (ORDER BY CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64))) < 1000
ORDER BY
  Unit
LIMIT
  10

Answer 1

要“声明”单元，您可以使用 CTE（常用 table 表达式）。

with time_lags as (
SELECT
  FORMAT_TIMESTAMP("%F", Timestamp) AS Date,
  FORMAT_TIMESTAMP("%R", Timestamp) AS Unit,
  LAG(FORMAT_TIMESTAMP("%R", Timestamp)) OVER (ORDER BY FORMAT_TIMESTAMP("%R", Timestamp)) AS CycleTimePreValue,
  CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64)-LAG(CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64)) OVER (ORDER BY CAST(FORMAT_TIMESTAMP("%s", Timestamp) AS INT64)) AS CycleTimePreValueSeconds,
FROM
  `wh-lr-sk.Quality.Retouren_IMPORT`
)
SELECT
    *
FROM time_lags
WHERE Unit < 1000
AND CycleTimePreValue IS NOT NULL
ORDER BY
  Unit
LIMIT
  10

要删除我刚刚添加的空值，请将 AND CycleTimePreValue IS NOT NULL 添加到 where 子句。

计算时间戳的差异并跳过 NULL

Calculate Difference of Timestamps and skip NULL

variables

timestamp

google-bigquery

difference