Oracle 查询具有特定列平均值的最新行?
Oracle query latest row with average for specific column?
具有以下数据集。我需要一些关于 sql 语句的帮助,它会根据 PING_DATE
和唯一的 PING_DESTINATION
和 PING_SOURCE
以及添加的 AVG
的列给我最新的行PING_AVG
最近 10 分钟内的所有行。
PING_DATE | PACKET_LOSS | PING_MIN | PING_AVG | PING_MAX | PING_SOURCE | PING_DESTINATION
-------------------------------------------------------------------------------------------------------
5/5/2015 12:58:18 PM | 0 | 68 | 68 | 72 | site1 | orange15
5/5/2015 12:58:43 PM | 0 | 68 | 71 | 76 | site1 | orange15
5/5/2015 12:59:11 PM | 0 | 68 | 68 | 72 | site1 | pear11
5/5/2015 1:09:47 PM | 0 | 68 | 70 | 76 | site1 | pear11
5/5/2015 1:43:59 PM | 0 | 68 | 69 | 72 | site1 | pear11
5/5/2015 1:45:41 PM | 0 | 68 | 69 | 72 | site1 | pear11
5/5/2015 2:03:43 PM | 0 | 68 | 68 | 72 | site1 | pear11
5/5/2015 3:01:53 PM | 0 | 68 | 68 | 72 | site1 | pear11
5/5/2015 3:02:05 PM | 0 | 68 | 69 | 72 | site1 | pear11
5/5/2015 3:00:59 PM | 20 | 68 | 68 | 68 | site1 | pear11
5/5/2015 3:01:07 PM | 0 | 68 | 68 | 72 | site1 | pear11
5/5/2015 3:01:14 PM | 0 | 68 | 70 | 72 | site1 | pear11
5/5/2015 12:46:55 PM | 3 | 3 | 3 | 3 | site1 | lemon1
查询结果:
PING_DATE | PACKET_LOSS | PING_MIN | PING_AVG | PING_MAX | PING_SOURCE | PING_DESTINATION | 10minavg
------------------------------------------------------------------------------------------------------------------
5/5/2015 12:58:43 PM | 0 | 68 | 71 | 76 | site1 | orange15 | 71
5/5/2015 3:01:14 PM | 0 | 68 | 70 | 72 | site1 | pear11 | 65
5/5/2015 12:46:55 PM | 3 | 3 | 3 | 3 | site1 | lemon1 | 3
像这样:
SELECT DISTINCT ping_source,
first_value(ping_date) over (partition by ping_source order by ping_date desc),
first_value(packet_loss) over (partition by ping_source order by ping_date desc)
-- ...
FROM data
CROSS JOIN
select avg(ping_avg) from data
where (sysdate - ping_date) * 24 * 60 < 10;
这是基于问题的直接查询。
根据示例输出进行了编辑。从现在开始的最后 10 分钟,在代码段“(i.latest_ping - interval '10' minute)”中使用 systemtimestamp 而不是 i.latest_ping。从 max_ping_time 开始的最后 10 分钟使用 i.latest_ping 作为源-目标对。
select
o.*,
(select avg(ping_avg) from ping_info a
where a.ping_source = i.ping_source
and a.ping_dest = i.ping_dest
and a.ping_date >= (systemtimestamp - interval '10' minute)
) last_10min_avg
from ping_info o,
(select ping_source, ping_dest, max(ping_date) latest_ping
from ping_info
group by ping_source, ping_dest) i
where o.ping_source = i.ping_source
and o.ping_dest = i.ping_dest
and o.ping_date = i.latest_ping;
对于 "last 10 minutes average" 是 "last 10 minutes in each group" 这是您要查找的查询:
with xyz as (
select X.*,
row_number() over (
partition by ping_destination, ping_source
order by ping_date desc
) as latest_row#,
avg(ping_avg) over (
partition by ping_destination, ping_source
order by ping_date asc
range between
interval '10' minute preceding
and current row
) as the_10_min_avg
from ping_table X
)
select *
from xyz
where latest_row# = 1
;
对于 "last 10 minutes average" 是 "from 10 minutes ago until now" 这是您要查找的查询:
with xyz as (
select X.*,
row_number() over (
partition by ping_destination, ping_source
order by ping_date desc
) as latest_row#,
avg(ping_avg) over (
partition by ping_destination, ping_source
) as the_10_min_avg
from ping_table X
where X.ping_date >= systimestamp - interval '10' minute
)
select *
from xyz
where latest_row# = 1
;
具有以下数据集。我需要一些关于 sql 语句的帮助,它会根据 PING_DATE
和唯一的 PING_DESTINATION
和 PING_SOURCE
以及添加的 AVG
的列给我最新的行PING_AVG
最近 10 分钟内的所有行。
PING_DATE | PACKET_LOSS | PING_MIN | PING_AVG | PING_MAX | PING_SOURCE | PING_DESTINATION
-------------------------------------------------------------------------------------------------------
5/5/2015 12:58:18 PM | 0 | 68 | 68 | 72 | site1 | orange15
5/5/2015 12:58:43 PM | 0 | 68 | 71 | 76 | site1 | orange15
5/5/2015 12:59:11 PM | 0 | 68 | 68 | 72 | site1 | pear11
5/5/2015 1:09:47 PM | 0 | 68 | 70 | 76 | site1 | pear11
5/5/2015 1:43:59 PM | 0 | 68 | 69 | 72 | site1 | pear11
5/5/2015 1:45:41 PM | 0 | 68 | 69 | 72 | site1 | pear11
5/5/2015 2:03:43 PM | 0 | 68 | 68 | 72 | site1 | pear11
5/5/2015 3:01:53 PM | 0 | 68 | 68 | 72 | site1 | pear11
5/5/2015 3:02:05 PM | 0 | 68 | 69 | 72 | site1 | pear11
5/5/2015 3:00:59 PM | 20 | 68 | 68 | 68 | site1 | pear11
5/5/2015 3:01:07 PM | 0 | 68 | 68 | 72 | site1 | pear11
5/5/2015 3:01:14 PM | 0 | 68 | 70 | 72 | site1 | pear11
5/5/2015 12:46:55 PM | 3 | 3 | 3 | 3 | site1 | lemon1
查询结果:
PING_DATE | PACKET_LOSS | PING_MIN | PING_AVG | PING_MAX | PING_SOURCE | PING_DESTINATION | 10minavg
------------------------------------------------------------------------------------------------------------------
5/5/2015 12:58:43 PM | 0 | 68 | 71 | 76 | site1 | orange15 | 71
5/5/2015 3:01:14 PM | 0 | 68 | 70 | 72 | site1 | pear11 | 65
5/5/2015 12:46:55 PM | 3 | 3 | 3 | 3 | site1 | lemon1 | 3
像这样:
SELECT DISTINCT ping_source,
first_value(ping_date) over (partition by ping_source order by ping_date desc),
first_value(packet_loss) over (partition by ping_source order by ping_date desc)
-- ...
FROM data
CROSS JOIN
select avg(ping_avg) from data
where (sysdate - ping_date) * 24 * 60 < 10;
这是基于问题的直接查询。
根据示例输出进行了编辑。从现在开始的最后 10 分钟,在代码段“(i.latest_ping - interval '10' minute)”中使用 systemtimestamp 而不是 i.latest_ping。从 max_ping_time 开始的最后 10 分钟使用 i.latest_ping 作为源-目标对。
select
o.*,
(select avg(ping_avg) from ping_info a
where a.ping_source = i.ping_source
and a.ping_dest = i.ping_dest
and a.ping_date >= (systemtimestamp - interval '10' minute)
) last_10min_avg
from ping_info o,
(select ping_source, ping_dest, max(ping_date) latest_ping
from ping_info
group by ping_source, ping_dest) i
where o.ping_source = i.ping_source
and o.ping_dest = i.ping_dest
and o.ping_date = i.latest_ping;
对于 "last 10 minutes average" 是 "last 10 minutes in each group" 这是您要查找的查询:
with xyz as (
select X.*,
row_number() over (
partition by ping_destination, ping_source
order by ping_date desc
) as latest_row#,
avg(ping_avg) over (
partition by ping_destination, ping_source
order by ping_date asc
range between
interval '10' minute preceding
and current row
) as the_10_min_avg
from ping_table X
)
select *
from xyz
where latest_row# = 1
;
对于 "last 10 minutes average" 是 "from 10 minutes ago until now" 这是您要查找的查询:
with xyz as (
select X.*,
row_number() over (
partition by ping_destination, ping_source
order by ping_date desc
) as latest_row#,
avg(ping_avg) over (
partition by ping_destination, ping_source
) as the_10_min_avg
from ping_table X
where X.ping_date >= systimestamp - interval '10' minute
)
select *
from xyz
where latest_row# = 1
;