Delete duplicates and keep one with condition in SQL Server - sql

Question

I am using SQL Server.
I have a table with the following design:
ID bigint
Number varchar(50)
Processed int
I have a lot of duplicates in the Number column
I want to delete all repeated Numbers, and keep the Number where processed=1
For Example if I have
Number --- Processed
111 --- 0
111 --- 0
111 --- 1
I want to delete all and keep the last one
Any help would be appreciated

Answer 1

Here is one method:
with todelete as (
select t.*,
row_number() over (partition by number order by processed desc) as seqnum
from table t
)
delete from todelete
where seqnum > 1;
The row_number() enumerates the rows, using the processed as a priority. The logic ensures that exactly one row remains, even if none have processed = 1.

Answer 2

If you are just trying to delete the rows where number equals 111 and processed does not equal 1 you can do:
delete from <table>
where
Number = 111 and
Processed <> 1
Assuming the ID is sequential and you want to keep the last row for each Number you can do:
delete from <table> t
left join (
select
MAX(ID) filter_ID
from <table>
group by
Number
) filter on
t.ID = filter.filter_ID
where
filter.filter_ID is null
to keep at least one row per Number giving priority to Processed = 1
delete from <table> t
left join (
select
ID
from (
select
ROW_NUMBER() OVER (
PARTITION BY
Number
ORDER BY
Processed DESC,
ID DESC
) last_R,
ID
from <table>
) filter
where
last_R = 1
) filter on
t.ID = filter.filter_ID
where
filter.filter_ID is null

Answer 3

Here is how I would approach this problem:
DECLARE #NUM VARCHAR(50)
DECLARE #TAB TABLE
(
NUMBER VARCHAR(50)
)
INSERT INTO #TAB
SELECT number, from <table> where processed = 0 GROUP BY number HAVING COUNT(number) > 1
DECLARE #IDToKEEP TABLE
(
id INT
)
WHILE (SELECT COUNT(*) FROM #TAB) > 0
BEGIN
SELECT TOP 1 #NUM = number FROM #TAB
INSERT INTO #IDToKEEP
SELECT TOP 1 id FROM <table> WHERE number = #NUM
DELETE FROM #TAB WHERE number = #NUM
END
DELETE FROM <table> WHERE processed = 0 AND number IN (SELECT number FROM #TAB) AND id NOT IN (SELECT id FROM #IDToKEEP)