Available in versions: Dev (3.19) | Latest (3.18) | 3.17 | 3.16 | 3.15 | 3.14 | 3.13 | 3.12 | 3.11 | 3.10 | 3.9

Lexical and logical SELECT clause order

Applies to ✅ Open Source Edition   ✅ Express Edition   ✅ Professional Edition   ✅ Enterprise Edition

SQL has a lexical and a logical order of SELECT clauses. The lexical order of SELECT clauses is inspired by the English language. As SQL statements are commands for the database, it is natural to express a statement in an imperative tense, such as "SELECT this and that!".

Logical SELECT clause order

The logical order of SELECT clauses, however, does not correspond to the syntax. In fact, the logical order is this:

The SQL Server documentation also explains this, with slightly different clauses:

  • FROM
  • ON
  • JOIN
  • WHERE
  • GROUP BY
  • WITH CUBE or WITH ROLLUP
  • HAVING
  • SELECT
  • DISTINCT
  • ORDER BY
  • TOP

As can be seen, databases have to logically reorder a SQL statement in order to determine the best execution plan.

Alternative syntaxes: LINQ, SLICK

Some "higher-level" abstractions, such as C#'s LINQ or Scala's SLICK try to inverse the lexical order of SELECT clauses to what appears to be closer to the logical order. The obvious advantage of moving the SELECT clause to the end is the fact that the projection type, which is the record type returned by the SELECT statement can be re-used more easily in the target environment of the internal domain specific language.

A LINQ example:

// LINQ-to-SQL looks somewhat similar to SQL
// AS clause    // FROM clause
From p          In db.Products

// WHERE clause
Where p.UnitsInStock <= p.ReorderLevel AndAlso Not p.Discontinued

// SELECT clause
Select p

A SLICK example:

// "for" is the "entry-point" to the DSL
val q = for {

    // FROM clause   WHERE clause
    c <- Coffees     if c.supID === 101

// SELECT clause and projection to a tuple
} yield (c.name, c.price)

While this looks like a good idea at first, it only complicates translation to more advanced SQL statements while impairing readability for those users that are used to writing SQL. jOOQ is designed to look just like SQL. This is specifically true for SLICK, which not only changed the SELECT clause order, but also heavily "integrated" SQL clauses with the Scala language.

For these reasons, the jOOQ DSL API is modelled in SQL's lexical order.

Feedback

Do you have any feedback about this page? We'd love to hear it!

The jOOQ Logo