SQL Subquery issue - Invalid object name

I have prepared the below "subquery" code. Error message: "Invalid object name 'group1-2'" is shown. May I know how to fix the issue.

select
  CASE
    WHEN reference_s = 'R1' THEN
      (select sum(DATEDIFF(second, datetime2_s,datetime6_s) ) from group1_2)
    WHEN reference_s = 'R2' THEN
      (select sum(DATEDIFF(second, datetime3_s,datetime6_s) ) from group1_2)
  END
  / (CAST(count(*) AS float)*60)
from
(
  select * from
  (
    select
      id,
      min(reference) as reference_s, min(datetime2) as datetime2_s,
      min(datetime3) as datetime3_s, max(datetime6) as datetime6_s,
      count(*) as type_s, sum(size) as size_s
    from dbo.testing
    group by id
  ) group1
  where type_s=2 and cntr_size_s=44
  UNION
  select * from
  (
    select
      id,
      min(reference) as reference_s, min(datetime2) as datetime2_s,
      min(datetime3) as datetime3_s, max(datetime6) as datetime6_s,
      count(*) as type_s, sum(size) as size_s
    from dbo.testing
    group by id
  ) group2
  where type_s=1
) group1_2
where datetime6_s IS NOT null
and datetime6_s >= '2023-03-01 00:00'
and datetime6_s <= '2023-03-30 23:59'

In your FROM clause you have a subquery the result of which you call group1_2. In your SELECT clause you have subqueries selecting from this. This is not allowed in SQL, because in the SELECT clause you look at a single row from the table (or an aggregated row of it). The SELECT clause does not know the whole table. If you want to define such an ad-hoc view that you can access multiple times in your query, you must make this a CTE (aka WITH clause).

But there are more flaws in your query:

• You have from (...) group1 where ... cntr_size_s = 44, but you don't select any cntr_size_s in group1.
• You have from (...) group2 where ... cntr_size_s = 44, but you don't select any cntr_size_s in group2.
• You have select case when reference_s = 'R1' ...end ... / cast(count(*) ...). So you access a column from group1_2 and an aggregation result. But there is no GROUP BY reference_s, which would give you an aggregation result per reference_s. So how could the data you are accessing here be aggregated to a single row (COUNT(*)) and not aggregated (reference_s) at the same time? This is not possible. GROUP BY reference_s could solve this syntactically, but I don't know whether this is the solution you are looking for.

Subquery which you used from the derived table is not allowed in SELECT part.
Apart from that few more logical issues are there in your query
If you want your query to run without any issue then try below

select
  CASE
    WHEN reference_s = 'R1' THEN
      sum(DATEDIFF(second, datetime2_s,datetime6_s))
    WHEN reference_s = 'R2' THEN
      sum(DATEDIFF(second, datetime3_s,datetime6_s) )
  END
  / (CAST(count(*) AS float)*60)
from
(
  select * from
  (
    select
      id,
      min(reference) as reference_s, min(datetime2) as datetime2_s,
      min(datetime3) as datetime3_s, max(datetime6) as datetime6_s,
      count(*) as type_s, sum(size) as size_s
    from dbo.testing
    group by id
  ) group1
  where type_s=2 and cntr_size_s=44
  UNION
  select * from
  (
    select
      id,
      min(reference) as reference_s, min(datetime2) as datetime2_s,
      min(datetime3) as datetime3_s, max(datetime6) as datetime6_s,
      count(*) as type_s, sum(size) as size_s
    from dbo.testing
    group by id
  ) group2
  where type_s=1
) group1_2
where datetime6_s IS NOT null
and datetime6_s >= '2023-03-01 00:00'
and datetime6_s <= '2023-03-30 23:59'
GROUP BY id,reference_s