Find all skipped numbers per group - sql-server

Question

I have this column in my table and I want to get all skipped transactions (which is a varchar)
SA1
SA3
SA50
SA999
I'm trying to get
SA2
SA4 to SA49
SA51 to SA998
EDIT: I've checked the actual data, found out that it's non-trailing zeroes.
I've done this with PHP. Was wondering if is there a way within SQL to do it?
This might sound dumb, but what I'm trying is to plot it like this (this returns null values)
WITH cte (id) AS (
SELECT ROW_NUMBER() OVER (ORDER BY s1.[object_id])
FROM sys.all_columns AS s1
CROSS JOIN sys.all_columns AS s2
),
table1 AS (
SELECT sNumber as id1 FROM Sales WHERE sNumber LIKE '%SI_'
),
table2 AS (
SELECT sNumber as id1 FROM Sales WHERE sNumber LIKE '%SI__'
),
table3 AS (
SELECT sNumber as id1 FROM Sales WHERE sNumber LIKE '%SI___'
)
SELECT
'SA' + RIGHT('' + CAST(t1.id AS VARCHAR(2)), 1) AS id_missing
FROM cte t1
LEFT JOIN table1 t2
ON t1.id = CAST(RIGHT(t2.id1, 1) AS INT)
WHERE
t1.id < (SELECT MAX(CAST(RIGHT(id1, 1) AS INT)) FROM yourTable1) AND
t2.id1 IS NULL
UNION ALL
SELECT
'SA' + RIGHT('' + CAST(t1.id AS VARCHAR(2)), 2) AS id_missing
FROM cte t1
LEFT JOIN table2 t2
ON t1.id = CAST(RIGHT(t2.id2, 2) AS INT)
WHERE
t1.id < (SELECT MAX(CAST(RIGHT(id2, 2) AS INT)) FROM yourTable2) AND
t2.id2 IS NULL
UNION ALL
SELECT
'SA' + RIGHT('' + CAST(t1.id AS VARCHAR(3)), 3) AS id_missing
FROM cte t1
LEFT JOIN table3 t2
ON t1.id = CAST(RIGHT(t2.id3, 3) AS INT)
WHERE
t1.id < (SELECT MAX(CAST(RIGHT(id3, 3) AS INT)) FROM yourTable3) AND
t2.id3 IS NULL

Answer 1

You may use a calendar table approach with a left join here:
WITH cte (id) AS (
SELECT ROW_NUMBER() OVER (ORDER BY s1.[object_id])
FROM sys.all_columns AS s1
CROSS JOIN sys.all_columns AS s2
)
SELECT
'SA' + RIGHT('000' + CAST(t1.id AS VARCHAR(3)), 3) AS id_missing
FROM cte t1
LEFT JOIN yourTable t2
ON t1.id = CAST(RIGHT(t2.id, 3) AS INT)
WHERE
t1.id < (SELECT MAX(CAST(RIGHT(id, 3) AS INT)) FROM yourTable) AND
t2.id IS NULL;
Demo
The idea is to generate a sequence of numbers covering the possible up to 1000 values which might appear in your current table as SAxxx. Then, we left join this calendar table to your current table, on the condition that the numeric portion of the id does not match. All such non matching SAxxx values are then retained in the result set.