在 Impala 中随时间 Window 的平均值 ... 结束（分区依据 ... 排序依据）

Question

我在 Impala 中有一个 Table，其中我的时间信息为 UnixTime，频率为 1 毫秒。我正在尝试获取 window 10 秒的 AVG()、MIN() 和 MAX()（但我不想修复它，可以是 20 秒、30 秒等）。

我正在使用子查询来做，但我没有得到正确的答案。以下是我在 Table 中的数据： Data in the Table

我正在使用以下子查询获取 window 10 秒的 AVG()、MIN() 和 MAX()。我正在使用 OVER (PARTITION BY ... ORDER BY) 但没有得到正确的结果。我的查询如下：

SELECT DISTINCT *
FROM
(SELECT ts,
last_value(Table1.val1) OVER (PARTITION BY Table1.ts ORDER BY Table1.ts rows between unbounded preceding and unbounded following) as val1,
AVG(Table1.val2) OVER (PARTITION BY Table1.ts ORDER BY Table1.ts rows between unbounded preceding and unbounded following) as val2
MIN(Table1.val3) OVER (PARTITION BY Table1.ts ORDER BY Table1.ts rows between unbounded preceding and unbounded following) as val2
MAX(Table1.val4) OVER (PARTITION BY Table1.ts ORDER BY Table1.ts rows between unbounded preceding and unbounded following) as val2
FROM (SELECT cast(cast(unix_timestamp(cast(ts/1000 as TIMESTAMP))/10 as bigint)*10 as TIMESTAMP) as ts , 
val1 as val1, 
val2 as val2, 
val3 as val3, 
val4 as val4 
FROM Sensor_Data.Table where unit='Unit1'  
and cast(ts/1000 as TIMESTAMP) BETWEEN '2020-11-29 22:30:00' and '2020-12-01 01:51:00') as Table1) as Table2
ORDER BY ts

我需要以下答案：

Time                    Val1        Val2        Val3        Val4
2020-11-29 22:30:00     last_value  AVG         MIN         MAX
2020-11-29 22:30:10     last_value  AVG         MIN         MAX
2020-11-29 22:30:20     last_value  AVG         MIN         MAX

谁能告诉我我的 Impala 查询有什么问题。

谢谢！！！

Answer 1

我认为您只需要聚合，而不是 window 函数：

SELECT cast(cast(unix_timestamp(cast(ts/1000 as TIMESTAMP))/10 as bigint)*10 as TIMESTAMP),
       AVG(val2) as val2,
       MIN(val3) as val3,
       MAX(val4) as val4
FROM Sensor_Data.Table
WHERE unit = 'Unit1' AND
      CAST(ts/1000 as TIMESTAMP) BETWEEN '2020-11-29 22:30:00' and '2020-12-01 01:51:00'
GROUP BY cast(cast(unix_timestamp(cast(ts/1000 as TIMESTAMP))/10 as bigint)*10 as TIMESTAMP)

在 Impala 中随时间 Window 的平均值 ... 结束（分区依据 ... 排序依据）

AVG over time Window in Impala ... OVER (PARTITION BY ... ORDER BY)

sql

database

partitioning

impala

qsqlquery