bigquery get highest possible steps group by col - sql

Question

i have question about counting row number based on a column iteration
my table looks like this
time | steps | name
13:02 | 0 | a
13:03 | 0 | a
13:04 | 1 | a
13:05 | 0 | a
13:07 | 1 | a
13:10 | 1 | a
13:12 | 2 | a
13:04 | 0 | b
13:06 | 0 | b
13:12 | 1 | b
13:14 | 2 | b
13:19 | 3 | b
13:14 | 0 | b
13:19 | 3 | b
from table above i want to get the highest possible steps made by name. but must meet these condition:
steps made by name must be sequential(ex: 0,1,2,3 return 0,1,2,3; 0,1,2,4 return 0,1,2)
each step must be sequential according to time
Select any value if there are more than 1 record is possible(ex: 0,1,1,2 return 0,ANY(1,1),2)
table i looking for is
time | steps | name
13:05 | 0 | a
13:07 | 1 | a
13:12 | 2 | a
13:06 | 0 | b
13:12 | 1 | b
13:14 | 2 | b
13:19 | 3 | b
Is there any way i can do this in bigquery?

Answer 1

First remove duplicates. Then identify the rows where the "next" step (by time) is what you expect.
The following almost works:
select t.*
from (select min(time) as time, steps, name,
lead(steps) over (partition by name order by min(time)) as next_step
from yourtable t
group by steps, name
) t
where next_step = step + 1;
However, you want the minimum set. For that, you also need for the row number to match. It turns out that that condition is sufficient:
select t.*
from (select min(time) as time, steps, name,
row_number() over (partition by name order by min(time)) as seqnum
from yourtable t
group by steps, name
) t
where step = seqnum - 1;