Question

I am in a process of adding non clustered indexes to SQL Azure DB and I wonder what is the difference between having multiple columns in a single non clustered index, compared to having multiple non clustered indexes with each having a single column in it?

Answer 1

the difference between having multiple columns in a single non clustered index, compared to having multiple non clustered indexes with each having a single column in it?

Consider a table T with a clucstered PK Id column, and additionally columns A, B, C.

A single nonclustered index containing A, B, and C could support fast lookup for queries such as:

WHERE A > @a
WHERE A = @a AND @b1 < B AND B < @b2
WHERE A = @a AND B = @b AND C < @c

but not

WHERE B = @b
WHERE C = @c

in both of which cases, we can’t do any better than a table scan

However, if we have multiple indexes, IX_A on A, etc:

WHERE A = @a
WHERE B = @b
WHERE C = @c

would all benefit from an index, and a composite query such as

WHERE A = @a AND B = @b AND C < @c

would have a small benefit too.