postgresql group by difference to previous row

Question

I have some sorted values (called point) that must be group together based on their difference to the previous row like in the following example.

In this example a new group is started when the difference of the point to the previous one is greater than 10.

point   group
  1     1
  2     1
  3     1
101     2
103     2
105     2
213     3
214     3
215     3

Using the lag window function I can calculate the difference between the points but I'm missing the part to build the groups like in the example above.

Any help is appreciated, thank you!

Answer 1

select point, sum(inc) over (order by point)+1 as group
from (
    select point, 
    (case when point-lag(point,1) over (order by point) > 1 then 1 else 0 end)  inc
    from whatevertable
) base
order by point

there may be a way to do it without a sub-select, but this is what I came up with.