SQL using If Not Null on a Concatenation

Here would be my suggestions:

PostgreSQL and other SQL databases where 'a' || NULL IS NULL, then use COALESCE:

SELECT firstname || COALESCE('-' || middlename, '') || '-' || surname ...

Oracle and other SQL databases where 'a' || NULL = 'a':

SELECT first name || DECODE(middlename, NULL, '', '-' || middlename) || '-' || surname...

I like to go for conciseness. Here it is not very interesting to any maintenance programmer whether the middle name is empty or not. CASE switches are perfectly fine, but they are bulky. I'd like to avoid repeating the same column name ("middle name") where possible.

As @Prdp noted, the answer is RDBMS-specific. What is specific is whether the server treats a zero-length string as being equivalent to NULL, which determines whether concatenating a NULL yields a NULL or not.

Generally COALESCE is most concise for PostgreSQL-style empty string handling, and DECODE (*VALUE*, NULL, ''... for Oracle-style empty string handling.

If you use Postgres, concat_ws() is what you are looking for:

SELECT concat_ws('-', Firstname, Middlename, Surname)  AS example_column
FROM example_table

SQLFiddle: http://sqlfiddle.com/#!15/9eecb7db59d16c80417c72d1e1f4fbf1/8812

To treat empty strings or strings that only contain spaces like NULL use nullif():

 SELECT concat_ws('-', Firstname, nullif(trim(Middlename), ''), Surname)  AS example_column
 FROM example_table