综合自拍亚洲综合图区高清,午夜电影街亚洲AV无码,无码av午夜福利

降采樣查詢是在時間維度的聚合查詢，是時序數據應用場景常用的降低采樣率的查詢。

引擎與版本

降采樣查詢僅支持時序引擎，且引擎版本需為3.4.15及以上版本。

說明

語法

select_sample_by_statement ::=  SELECT ( select_clause | '*' )
                                FROM table_identifier
                                WHERE where_clause
                                SAMPLE BY time_interval [ OFFSET offset_interval ] [ FILL fill_option ]
select_clause              ::=  selector [ AS identifier ] ( ',' selector [ AS identifier ] )
selector                   ::=  tag_identifier | time | function_identifier '(' field_identifier [ ',' function_args ] ')'
where_clause               ::=  relation ( AND relation )* (OR relation)*
relation                   ::=  ( field_identifier | tag_identifier ) operator term
operator                   ::=  '=' | '<' | '>' | '<=' | '>=' | '!=' | IN | CONTAINS | CONTAINS KEY
time_interval              ::= interval units | 0
offset_interval            ::= interval units

降采樣支持的聚合函數列表

SAMPLE BY是針對每條獨立的時間線（關于時間線的概念，請參見數據模型）上的降采樣操作。

SAMPLE BY支持的函數列表：

函數	說明
SUM	每個指定時間窗口內求和，具體請參見SUM函數。
AVG	每個指定時間窗口內求均值，具體請參見AVG函數。
COUNT	每個指定時間窗口內值個數，具體請參見COUNT函數。
MIN	每個指定時間窗口內最小值，具體請參見MIN函數。
MAX	每個指定時間窗口內最大值，具體請參見MAX函數。
FIRST	每個指定時間窗口內第一個值，具體請參見FIRST函數。
LAST	每個指定時間窗口內最后值，具體請參見LAST函數。
PERCENTILE	每個指定時間窗口內求百分位，具體請參見PERCENTILE函數。
LATEST	整個時間區間最新值，具體請參見LATEST函數。
RATE	與前一行對應值的變化率，具體請參見RATE函數。
DELTA	與前一行對應值的差，具體請參見DELTA函數。

示例

SELECT查詢指定的tag列無需指定降采樣函數，其他field列必須指定降采樣函數。

假定查詢的sensor表數據如下：

SELECT * FROM sensor;

返回結果如下：

+-----------+----------+---------------------------+-------------+-----------+
| device_id |  region  |           time            | temperature | humidity  |
+-----------+----------+---------------------------+-------------+-----------+
| F07A1260  | north-cn | 2021-01-01T09:00:00+08:00 | 0.000000    | 9.000000  |
| F07A1260  | north-cn | 2021-01-01T12:01:00+08:00 | 1.000000    | 45.000000 |
| F07A1260  | north-cn | 2021-01-01T14:03:00+08:00 | 2.000000    | 46.000000 |
| F07A1260  | north-cn | 2021-01-01T20:00:00+08:00 | 10.000000   | 47.000000 |
| F07A1261  | north-cn | 2021-02-10T12:00:30+08:00 | 3.000000    | 40.000000 |
| F07A1261  | north-cn | 2021-03-01T12:01:00+08:00 | 4.000000    | 41.000000 |
| F07A1261  | north-cn | 2021-03-08T12:08:00+08:00 | 5.000000    | 42.000000 |
| F07A1261  | north-cn | 2021-05-01T13:00:00+08:00 | 6.000000    | 43.000000 |
+-----------+----------+---------------------------+-------------+-----------+

降采樣與子查詢示例

說明

降采樣查詢中不支持嵌套子查詢，但可以作為子查詢被其他查詢嵌套。

示例1：默認UTC對齊降采樣，時間線按照8h時間區間聚合分別求count。

SELECT device_id,region,time,count(humidity) AS count_humidity FROM sensor WHERE device_id='F07A1260' sample by 8h;

返回結果如下：

+-----------+----------+---------------------------+----------------+
| device_id |  region  |           time            | count_humidity |
+-----------+----------+---------------------------+----------------+
| F07A1260  | north-cn | 2021-01-01T08:00:00+08:00 | 3              |
| F07A1260  | north-cn | 2021-01-01T16:00:00+08:00 | 1              |
+-----------+----------+---------------------------+----------------+

示例2：默認UTC對齊降采樣，指定窗口offset偏移，時間線按照8h時間區間聚合，開始窗口偏移3h，求count。

SELECT device_id,region,time,count(humidity) AS count_humidity FROM sensor WHERE device_id='F07A1260' sample by 8h offset 3h;

返回結果如下：

+-----------+----------+---------------------------+----------------+
| device_id |  region  |           time            | count_humidity |
+-----------+----------+---------------------------+----------------+
| F07A1260  | north-cn | 2021-01-01T03:00:00+08:00 | 1              |
| F07A1260  | north-cn | 2021-01-01T11:00:00+08:00 | 2              |
| F07A1260  | north-cn | 2021-01-01T19:00:00+08:00 | 1              |
+-----------+----------+---------------------------+----------------+

示例3: 默認UTC對齊降采樣，對齊到當地時間0點（例如東8區時間），時間線按照24h時間區間聚合，開始窗口偏移16h，求count。

SELECT device_id,region,time,count(humidity) AS count_humidity FROM sensor WHERE device_id='F07A1260' sample by 24h offset 16h

返回結果如下：

+-----------+----------+---------------------------+----------------+
| device_id | region   | time                      | count_humidity |
+-----------+----------+---------------------------+----------------+
| F07A1260  | north-cn | 2021-01-01T00:00:00+08:00 | 4              |
+-----------+----------+---------------------------+----------------+

示例4：sample by目前不支持與group by、limit offset、order by配合，但可以配合子查詢使用。

SELECT device_id, max(avg_humidity) AS max_humidity  FROM (SELECT device_id,region,time,avg(humidity) AS avg_humidity FROM sensor sample by 8h) group by device_id;

返回結果如下：

+-----------+--------------+
| device_id | max_humidity |
+-----------+--------------+
| F07A1261  | 43.000000    |
| F07A1260  | 47.000000    |
+-----------+--------------+

示例5：配合limit offset限制結果條數。

SELECT device_id,region, avg_humidity FROM (select device_id,region,time,avg(humidity) AS avg_humidity FROM sensor sample by 8h) limit 1 offset 1;

返回結果如下：

+-----------+----------+--------------+
| device_id |  region  | avg_humidity |
+-----------+----------+--------------+
| F07A1261  | north-cn | 40.000000    |
+-----------+----------+--------------+

降采樣窗口插值示例

示例1：固定值插值。

SELECT * from (select device_id,region,time, avg(humidity) AS humidity FROM sensor WHERE device_id='F07A1260' sample by 2h fill 1) order by device_id;

返回結果如下：

+-----------+----------+---------------------------+-----------+
| device_id |  region  |           time            | humidity  |
+-----------+----------+---------------------------+-----------+
| F07A1260  | north-cn | 2021-01-01T08:00:00+08:00 | 9.000000  |
| F07A1260  | north-cn | 2021-01-01T10:00:00+08:00 | 1.000000  |
| F07A1260  | north-cn | 2021-01-01T12:00:00+08:00 | 45.000000 |
| F07A1260  | north-cn | 2021-01-01T14:00:00+08:00 | 46.000000 |
| F07A1260  | north-cn | 2021-01-01T16:00:00+08:00 | 1.000000  |
| F07A1260  | north-cn | 2021-01-01T18:00:00+08:00 | 1.000000  |
| F07A1260  | north-cn | 2021-01-01T20:00:00+08:00 | 47.000000 |
+-----------+----------+---------------------------+-----------+

示例2：后值插值。

SELECT * from (select device_id,region,time,avg(humidity) AS humidity FROM sensor WHERE device_id='F07A1260' sample by 2h fill after) order by device_id;

返回結果如下：

+-----------+----------+---------------------------+-----------+
| device_id |  region  |           time            | humidity  |
+-----------+----------+---------------------------+-----------+
| F07A1260  | north-cn | 2021-01-01T08:00:00+08:00 | 9.000000  |
| F07A1260  | north-cn | 2021-01-01T10:00:00+08:00 | 45.000000 |
| F07A1260  | north-cn | 2021-01-01T12:00:00+08:00 | 45.000000 |
| F07A1260  | north-cn | 2021-01-01T14:00:00+08:00 | 46.000000 |
| F07A1260  | north-cn | 2021-01-01T16:00:00+08:00 | 47.000000 |
| F07A1260  | north-cn | 2021-01-01T18:00:00+08:00 | 47.000000 |
| F07A1260  | north-cn | 2021-01-01T20:00:00+08:00 | 47.000000 |
+-----------+----------+---------------------------+-----------+

降采樣后再轉換示例

示例1：先按照2h窗口avg降采樣，然后對降采樣后的數據計算rate斜率。

SELECT device_id,region,time,rate(avg(humidity)) AS rate_humidity FROM sensor WHERE device_id='F07A1260' sample by 2h;

返回結果如下：

+-----------+----------+---------------------------+---------------+
| device_id |  region  |           time            | rate_humidity |
+-----------+----------+---------------------------+---------------+
| F07A1260  | north-cn | 2021-01-01T12:00:00+08:00 | 0.002500      |
| F07A1260  | north-cn | 2021-01-01T14:00:00+08:00 | 0.000139      |
| F07A1260  | north-cn | 2021-01-01T20:00:00+08:00 | 0.000046      |
+-----------+----------+---------------------------+---------------+

示例2：先按照2h窗口降采樣，然后對降采樣后的數據計算差值delta。

SELECT device_id,region,time,delta(avg(humidity)) AS humidity FROM sensor WHERE device_id='F07A1260' sample by 2h;

返回結果如下：

+-----------+----------+---------------------------+-----------+
| device_id |  region  |           time            | humidity  |
+-----------+----------+---------------------------+-----------+
| F07A1260  | north-cn | 2021-01-01T12:00:00+08:00 | 36.000000 |
| F07A1260  | north-cn | 2021-01-01T14:00:00+08:00 | 1.000000  |
| F07A1260  | north-cn | 2021-01-01T20:00:00+08:00 | 1.000000  |
+-----------+----------+---------------------------+-----------+

插值

降采樣先把所有時間線按照指定時間窗口切分，并把每個降采樣區間內的數據做一次運算，降采樣后如果某個精度區間沒有值，插值可以指定在這個時間點填充具體的值。比如某條時間線降采樣后的時間戳為：t+0, t+20, t+30，此時如果不指定插值，只有3個值，如果指定了插值為1，此時間線會有4個值，其中t+10時刻的值為1。

插值函數表：

Fill Policy	填充值
none	默認行為，不填值。
zero	固定填入0。
linear	線性填充值。
previous	之前的一個值。
near	鄰近的一個值。
after	之后的一個值。
fixed	用指定的一個固定填充值。

日本熟妇hd丰满老熟妇,中文字幕一区二区三区在线不卡 ,亚洲成片在线观看,免费女同在线一区二区

引擎與版本

語法

降采樣支持的聚合函數列表

示例

降采樣與子查詢示例

降采樣窗口插值示例

降采樣后再轉換示例

插值