SQL query help to calculate max - sql

Question

I have a question regarding a query I need to do in SQL (i use BQ).
I have this table:
train_no, wagon_no, weight, length, date, startpoint(km), endpoint(km)
1, 123, 1000, 20, 20190101, 0, 7
1, 234, 2000, 20, 20190101, 1, 2
1, 345, 3000, 30, 20190101, 1, 5
1, 456, 1000, 40, 20190101, 1, 6
2, 987, 1000, 10, 20190101, 0, 8
2, 876, 2000, 20, 20190101, 1, 2
2, 765, 3000, 20, 20190101, 1, 5
2, 654, 1000, 20, 20190101, 1, 6
The table shows two trains with wagons. Per wagon we see at what point the wagon was added on the train. So for train no 1 we see that wagon 234 was included on the train from startpoint=1 (kilometer 1) to endpoint=2 (kilometer 2) then it was removed from the train. We also see that max endpoint is 7 for train_no =1 so max travelled distance for the train is 7 km.
The total train length and weight varies during the distance and I would like to calculate the maximum length and maximum weight reached during the distance. How can I do this in SQL?
Any suggestion is appreciated.
Edit:
Adding a pic to clearify what Im looking for.
As you can see in attached pic Train_no =1 have a max weight between point 1 and 2. Total weight is 7000 and is the total of all wagons in the train at that specific distance. Also, total length is 110 which is the total length of the all wagons added together.

Answer 1

Below is for BigQuery Standard SQL
#standardSQL
WITH temp AS (
SELECT train_no, dt, MIN(startpoint) startpoint, MAX(endpoint) endpoint
FROM `project.dataset.table`
GROUP BY train_no, dt
)
SELECT train_no, dt, MAX(wagons) max_wagons, MAX(total_weight) AS max_total_weight, MAX(total_len) max_total_len
FROM (
SELECT train_no, dt, point, COUNT(wagon_no) wagons, SUM(weight) total_weight, SUM(len) total_len
FROM temp, UNNEST(GENERATE_ARRAY(startpoint, endpoint)) point
LEFT JOIN `project.dataset.table` t
USING(train_no, dt)
WHERE point >= t.startpoint AND point < t.endpoint
GROUP BY train_no, dt, point
)
GROUP BY train_no, dt
If to apply to sample data from your question as in example below
#standardSQL
WITH `project.dataset.table` AS (
SELECT 1 train_no, 123 wagon_no, 1000 weight, 20 len, '20190101' dt, 0 startpoint, 7 endpoint UNION ALL
SELECT 1, 234, 2000, 20, '20190101', 1, 2 UNION ALL
SELECT 1, 345, 3000, 30, '20190101', 1, 5 UNION ALL
SELECT 1, 456, 1000, 40, '20190101', 1, 6 UNION ALL
SELECT 2, 987, 1000, 10, '20190101', 0, 8 UNION ALL
SELECT 2, 876, 2000, 20, '20190101', 1, 2 UNION ALL
SELECT 2, 765, 3000, 20, '20190101', 1, 5 UNION ALL
SELECT 2, 654, 1000, 20, '20190101', 1, 6
), temp AS (
SELECT train_no, dt, MIN(startpoint) startpoint, MAX(endpoint) endpoint
FROM `project.dataset.table`
GROUP BY train_no, dt
)
SELECT train_no, dt, MAX(wagons) max_wagons, MAX(total_weight) AS max_total_weight, MAX(total_len) max_total_len
FROM (
SELECT train_no, dt, point, COUNT(wagon_no) wagons, SUM(weight) total_weight, SUM(len) total_len
FROM temp, UNNEST(GENERATE_ARRAY(startpoint, endpoint)) point
LEFT JOIN `project.dataset.table` t
USING(train_no, dt)
WHERE point >= t.startpoint AND point < t.endpoint
GROUP BY train_no, dt, point
)
GROUP BY train_no, dt
result is
Row train_no dt max_wagons max_total_weight max_total_len
1 1 20190101 4 7000 110
2 2 20190101 4 7000 70

Answer 2

The following query returns the the length at each KM marker sorted in descending fashion for each train.
with data as (
select 1 as train_no, 123 as wagon_no, 1000 as weight, 20 as length, 20190101 as date, 0 as startpoint, 7 as endpoint union all
select 1, 234, 2000, 20, 20190101, 1, 2 union all
select 1, 345, 3000, 30, 20190101, 1, 5 union all
select 1, 456, 1000, 40, 20190101, 1, 6 union all
select 2, 987, 1000, 10, 20190101, 0, 8 union all
select 2, 876, 2000, 20, 20190101, 1, 2 union all
select 2, 765, 3000, 20, 20190101, 1, 5 union all
select 2, 654, 1000, 20, 20190101, 1, 6
),
km_array as (
select * from unnest(generate_array(0,10)) km
),
joined as (
select *
from km_array
cross join data
where km between startpoint and endpoint
),
train_length_at_each_km as (
select
km,
train_no,
sum(length) as length
from joined
group by 1,2
)
select
train_no, length, km
from train_length_at_each_km
order by train_no, length desc
Getting maximum weight would use similar logic as the train_length_at_each_km CTE.