How to not display an item in select query? - sql

Question

I feel a little stupid asking this because I feel like this is very easy, but for some reason I'm not able to update a query to not select a specific item based on two criteria.
Let's say I have data like this:
ID Name Variant Count1
110 Bob Type1 0
110 Bob Type2 1
120 John Type1 1
So as you can see we have two BOB rows with same ID but different variant (type1 and type2). I want to be able to only see one of the Bob's.
Desired result:
110 Bob Type2
120 John Type1
So what I've been doing is something like
Select ID, Name, Variant, sum(count1) from tbl1
where (id not in (110) and Variant <> 'type1')
Group by Id,name,variant
Please don't use COUNT as a criteria, because in my example it just so happens that Count=0 for the row that I don't want to see. It can vary.
I have many rows where I can have multiple instances of the same id with a variety of different VARIANTS. I'm looking to exclude certain instances of ID based on Variant value
UPDATE:
It has nothing to do with latest variant, it has to do with a specific variant. So I'm just looking to basically be able to use a clause where i used the ID and VARIANT, in order to remove that particular row.

Answer 1

Aggregating (grouping) the data like you're doing is one way to do it, although the where condition is a little overkill. If all you want to do is see the unique combinations of ID and Name, then another approach is just to use the "distinct" statement.
select distinct Id, Name
from tbl1
If you always want to see data from a specific Variant then just include that condition in your where clause and you don't need to worry about using distinct or aggregates.
select *
from tbl1
where Variant = 'Type 1'
If you always want to see the record associated with the latest Variant, then you can use a window function to do so.
select a.Id, a.Name, a.Variant
from
(
select *, row_number() over (partition by Id order by Variant desc) as RowRank
from tbl1
) a
where RowRank = 1
;
If there is not a predictable pattern for exclusion then you will have to maintain an exclusion list. It's not ideal but if you want to maintain this in the SQL itself then you could have a query like the one below.
select *
from tbl1
-- Define rows to exlcude
where not (Id = 110 and Variant = 'Type 1') -- Your example
and not (Id = 110 and Variant = 'Type 3') -- Theoretical example
;
A better solution would be to create an exclusion reference table to maintain all exclusions within. Then you could simply negative join to that table to retrieve your desired results.

Answer 2

Have you considered using an exclusion table where you can place the ID and Variant combinations that you want to exclude? ( I just used temp tables for this example, you can always use user tables so your exclusion table will always be available)
Here is an example of what I mean based on your example:
if object_id('tempdb..#temp') is not null
drop table #temp
create table #temp (
ID int,
Name varchar(20),
Variant varchar(20),
Count1 int
)
if object_id('tempdb..#tempExclude') is not null
drop table #tempExclude
create table #tempExclude (
ID int,
Variant varchar(20)
)
insert into #temp values
(110,'Bob','Type1',0),
(110,'Bob','Type2',1),
(120,'John','Type1',1),
(120,'John','Type2',1),
(120,'John','Type2',1),
(120,'John','Type2',1),
(120,'John','Type3',1)
insert into #tempExclude values (110,'Type1')
select
t.ID,
t.Name
,t.Variant
,sum(t.Count1) as TotalCount
from
#temp t
left join
#tempExclude te
on t.ID = te.ID
and t.Variant = te.Variant
where
te.id is null
group by
t.ID,
t.Name
,t.Variant
Here are the results:

Answer 3

I think the logic you want is something like:
Select ID, Name, Variant, sum(count1)
from tbl1
where not (id = 110 and variant = 'type1')
Group by Id, name, variant;
For the second condition, just keep adding:
where not (id = 110 and variant = 'type1') and
not (id = 314 and variant = 'popsicle')
You can also express this using a list of exclusions:
select t.ID, Name, t.Variant, sum(t.count1)
from tbl1 t left join
(values (111, 'type1'),
(314, 'popsicle')
) v(id, excluded_variant)
on t.id = v.id and
t.variant = v.excluded_variant
where v.id is not null -- doesn't match an exclusion criterion
group by Id, name, variant;