Delete records with condition

Question

I have a table over which I have built an ETL service. Goods records (arrival / departure) go to the table. I have done that my table will be erased. When the item identifier arrives in the database for the second time, both records are deleted.

label   cost   time
x2       29    14/5/2020 01:00:00
x3       20    14/5/2020 01:02:00
x2       29    15/5/2020 03:12:02

Now ETL service remove records (every 30s):

label   cost   time
x3       20    14/5/2020 01:02:00

I delete it using the function:

with todelete as (
      select *, count(*) over (partition by label) as cnt, ROW_NUMBER() over (partition by label order by time DESC) as r_number
      from Table1
     )
delete from todelete
    where cnt >= 2 

I just need to make a small change right now. When, for example, during those 30 seconds I get the second record in the table, but also the third that the goods have arrived again.

        label   cost   time
        x2       29    14/5/2020 01:00:00
        x3       20    14/5/2020 01:02:00
        x2       29    15/5/2020 03:12:02
        x2       29    15/5/2020 03:12:22

In this case, I would lose the last record of the arrival of the goods. So I need to adjust the delete function. I've tried this, but it doesn't work properly.

with todelete as (
      select *
              , count(*) over (partition by label) as cnt
              , ROW_NUMBER() over (partition by label order by time DESC) as r_number
      from Table1
     )
delete from todelete
    where (cnt = 2 OR cnt = 4) OR (cnt = 3 AND r_number > 1);

Would anyone help me to modify that function please.

UPDATE: Exactly table: Step1:

label   cost   time
x2       29    14/5/2020 01:00:00
x3       20    14/5/2020 01:02:00
x2       29    15/5/2020 03:12:02

Step2: Delete row with function:

with todelete as (
      select *, count(*) over (partition by label) as cnt, ROW_NUMBER() over (partition by label order by time DESC) as r_number
      from Table1
     )
delete from todelete
    where cnt >= 2 

My goal is described above, but one more thing that affects is when it comes to moving with a change in price like this:

  label   cost   time
    x2       29    14/5/2020 01:00:00
    x3       20    14/5/2020 01:02:00
    x2       30    15/5/2020 03:12:22

And then ETL deleted the old record and left the one with the new price.

        label   cost   time
        x3       20    14/5/2020 01:02:00
        x2       30    15/5/2020 03:12:22

This fact complicates it even more.

Answer 1

why don't you try something like this.

Explanation:

Notice that I changed from DESC to ASC in the row_number() function's order by clause.

with this I delete the first arriving rows till the rows are even. if Cnt=3 then (cnt/2)*2=2 as cnt is integer(note integer/2 gives integer not a fraction)

when cnt=4 then (cnt/2)*2 gives back 4.

with todelete as (
      select *, count(*) over (partition by label) as cnt, ROW_NUMBER() over (partition by label order by time ASC) as r_number
      from Table1
     )
delete from todelete 
    where cnt > 1 and r_number between 1 and (cnt/2)*2