在SQL中对列中的序列号进行编号

Question

I want to give a number to the value based on value in above row such that when the sequence breaks, it should again start from 1 otherwise should keep on increment the number.

The query is :

select'30300001' as lst union all
select'30300002' union all
select'30300003' union all
select'30300004' union all
select'30300001' union all
select'30300006' union all
select'30300007' union all
select'30300008' union all
select'30300009'

And the output I want to be as:

select'30300001' as lst,1 as rnk union all
select'30300002',2 union all
select'30300003',3 union all
select'30300004',4 union all
select'30300001',1 union all
select'30300006',1 union all
select'30300007',2 union all
select'30300008',3 union all
select'30300009',4 

I tried it with row_number and rank functions but could not get the required output. How can I get the desired result?

Answer 1

If you have an orderingid (of some sort), then you can use the difference between your column and a column with a sequence. This identifies a group, which can then be used with other window functions:

select q.*,
       row_number() over (partition by grp order by orderingid)
from (select q.*,
             (lst - row_number() over (order by orderingid)) as grp
      from query q
     ) q;

Note: this assumes that lst is actually a number or a value readily converted to a number.