Find the lowest non-contiguous value - sql

Question

I was asked to write a T-SQL statement that will find the lowest unused value of MyId in the sequence below (i.e. in this case the result should be 3):
DECLARE #MyTable TABLE (MyId INT);
INSERT INTO #MyTable(MyId) VALUES(1),(2),(4),(5);

Answer 1

;With CTE
AS
(
SELECT * , ROW_NUMBER() OVER (ORDER BY MyID ASC) AS RN
FROM #MyTable
)
SELECT TOP 1 rn
FROM CTE
WHERE Rn <> MyId
ORDER BY MyId ASC

Answer 2

Here is how I would probably answer that question (though it is impossible for anyone here to know exactly what the interviewer was after).
First, you can easily generate a sequence of contiguous numbers from existing tables or views in any SQL Server system. For this, let's use master..spt_values (which will cover a sequence of about 2000 values, depending on version):
SELECT TOP (5) n = number + 1
FROM master.dbo.spt_values
WHERE type = N'P'
ORDER BY number;
Results:
n
------
1
2
3
4
5
Now, you don't know in advance that you need 5, so you can determine the number you need by taking the min and max from the table:
DECLARE #min INT, #max INT;
SELECT #min = MIN(MyId), #max = MAX(MyId) FROM #MyTable;
Now you can get the exact set you need (since it may not always start at 1):
SELECT TOP (#max-#min+1) number
FROM master.dbo.spt_values
WHERE number >= #min AND type = N'P'
ORDER BY number;
Now, finally, we can perform a left anti-semi-join to find the first value that exists in our contiguous set but not in the table:
;WITH x AS
(
SELECT TOP (#max-#min+1) number
FROM master.dbo.spt_values
WHERE number >= #min AND type = N'P'
ORDER BY number
)
SELECT MIN(number) FROM x
WHERE NOT EXISTS
(SELECT 1 FROM #MyTable WHERE MyId = x.number);
If you need more than 2000 values, you can use other things like sys.all_columns and if that isn't enough you can CROSS JOIN multiple tables. See http://www.sqlperformance.com/generate-a-set-1, http://www.sqlperformance.com/generate-a-set-2 and http://www.sqlperformance.com/generate-a-set-3.
Of course, if you know the sequence should always start at 1, rather than the minimum value in the table, then the other answers are slightly simpler. This caters to the case where the set doesn't necessarily start with 1, and you don't care about "missing" values that are below the minimum value.

Answer 3

with cte as (
select MyId, row_number over (order by MyId asc) RowId
from #MyTable
)
select top 1 c1.MyId + 1 FirstMissingMyId
from cte c1
join cte c2
on c1.RowId + 1 = c2.RowId
where c1.MyId + 1 <> c2.MyId
order by c1.MyId asc

Answer 4

The idea is to find first non-sequential number for a sequence that strats with (assumingly) with 1. You can do this by comparing to sequential numbers, generated by ROW_NUMBER() function
SELECT TOP(1) RN FROM
(SELECT MyID, ROW_NUMBER() OVER (ORDER BY MyId) AS RN FROM #MyTable) MT
WHERE MyID <> RN
ORDER BY MyID ASC
Demo: http://sqlfiddle.com/#!3/c1a90/2