I have two numbers as input from the user, like for example 1000
and 1050
.
How do I generate the numbers between these two numbers, using
I know I'm 4 years too late, but I stumbled upon yet another alternative answer to this problem. The issue for speed isn't just pre-filtering, but also preventing sorting. It's possible to force the join-order to execute in a manner that the Cartesian product actually counts up as a result of the join. Using slartidan's answer as a jump-off point:
WITH x AS (SELECT n FROM (VALUES (0),(1),(2),(3),(4),(5),(6),(7),(8),(9)) v(n))
SELECT ones.n + 10*tens.n + 100*hundreds.n + 1000*thousands.n
FROM x ones, x tens, x hundreds, x thousands
ORDER BY 1
If we know the range we want, we can specify it via @Upper and @Lower. By combining the join hint REMOTE along with TOP, we can calculate only the subset of values we want with nothing wasted.
WITH x AS (SELECT n FROM (VALUES (0),(1),(2),(3),(4),(5),(6),(7),(8),(9)) v(n))
SELECT TOP (1+@Upper-@Lower) @Lower + ones.n + 10*tens.n + 100*hundreds.n + 1000*thousands.n
FROM x thousands
INNER REMOTE JOIN x hundreds on 1=1
INNER REMOTE JOIN x tens on 1=1
INNER REMOTE JOIN x ones on 1=1
The join hint REMOTE forces the optimizer to compare on the right side of the join first. By specifying each join as REMOTE from most to least significant value, the join itself will count upwards by one correctly. No need to filter with a WHERE, or sort with an ORDER BY.
If you want to increase the range, you can continue to add additional joins with progressively higher orders of magnitude, so long as they're ordered from most to least significant in the FROM clause.
Note that this is a query specific to SQL Server 2008 or higher.