在PostgreSQL中找到一个集合之间的所有行

Question

[我有一个名为tc_fuel的表，该表正在从GPS车辆接收所有与燃料相关的数据，我想获取最后的“燃料箱使用情况”，以计算整个燃料箱的MPG，但是当燃料箱已满时的读数（ 100）在2或3或更多行之后重复一些，所以我剩下2个彼此相邻的100值，我希望能够获得最后一个“填充”开始和结束ID。

到目前为止我所拥有的：

SELECT
    "tc_fuel".deviceid,
    "tc_fuel"."id",
    "tc_fuel".fuel,
    "tc_fuel"."fuelUsed",
    "tc_fuel"."fuelUsed"
FROM "tc_fuel"
WHERE fuel=100
    AND deviceid=19
ORDER BY ID
DESC LIMIT 2

然后，我进入PHP检查ID差是否超过100条记录，以检查燃料值是否彼此相邻，但这比我想知道是否有更好的方法要多得多。

例如，这辆车从满油箱开始，然后下降到6％的油箱并充满了油箱，我希望能够获取最后一个油箱的所有数据。

id    | deviceId  | fuel
------+-----------+-------
1     | 19        | 100  <-- This should be starting point 
2     | 19        | 97  
3     | 19        | 100  
4     | 19        | 96
5     | 19        | 94
6     | 19        | .... (keeps dropping)
7     | 19        | 33
8     | 19        | 31
9     | 19        | 30
10    | 19        | ....
11    | 19        | 6
12    | 19        | 5
13    | 19        | 6    <-- This should be end point (will flag this id as processed)
14    | 19        | 100  <-- Starts all over again in next iteration of the php script
15    | 19        | 99
16    | 19        | 98
17    | 19        | 100
18    | 19        | 99
19    | 19        | 97
20    | 19        | 96
21    | 19        | ....

Answer 1

SELECT min(id) AS first_id,
       max(id) AS last_id,
       deviceid
FROM (SELECT id, deviceid, fuel,
             count(*) FILTER (WHERE refilled)
                OVER (PARTITION BY deviceid ORDER BY id DESC) AS filling
      FROM (SELECT id, deviceid, fuel,
                   fuel < lead(luel, 1, 0)
                             OVER (PARTITION BY deviceid ORDER BY id) AS before_fill
            FROM tc_fuel
           ) AS refill
     ) AS fills
WHERE filling = 1
GROUP BY deviceid;

Answer 2

“填充”的定义有点含糊。我假设当燃油值增加超过50