Available in versions: Dev (3.20) | Latest (3.19) | 3.18 | 3.17 | 3.16 | 3.15 | 3.14 | 3.13 | 3.12 | 3.11 | 3.10

This documentation is for the unreleased development version of jOOQ. Click on the above version links to get this documentation for a supported version of jOOQ.

Synthetic SQL clauses

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

Most of the previously mentioned SQL clauses have a native representation in at least one of jOOQ's supported SQL dialects. For example, when a function like LPAD() is unavailable, jOOQ produces an equivalent expression for it:

-- MySQL (native support)
lpad('a', 10, ' ')

-- SQL Server (emulation)
(replicate(' ', (10 - len('a'))) + 'a')
// In Java



lpad("a", 10, " ")

However, since a lot of SQL is emulated for dialect compatibility, nothing prevents jOOQ from supporting synthetic SQL clauses that do not have any native representation anywhere.

An example for this is the quantified like predicate, introduced in jOOQ 3.12, which would be really useful in any database:

(TITLE     LIKE '%abc%'  OR TITLE     LIKE '%def%')
(TITLE NOT LIKE '%abc%'  OR TITLE NOT LIKE '%def%')
(TITLE     LIKE '%abc%' AND TITLE     LIKE '%def%')
(TITLE NOT LIKE '%abc%' AND TITLE NOT LIKE '%def%')
BOOK.TITLE.like(any("%abc%", "%def%"))
BOOK.TITLE.notLike(any("%abc%", "%def%"))
BOOK.TITLE.like(all("%abc%", "%def%"))
BOOK.TITLE.notLike(all("%abc%", "%def%"))

In this section, we briefly list most such synthetic SQL clauses, which are available both through the jOOQ API, and through the jOOQ parser, yet they do not have a native representation in any dialect.

  • Implicit JOIN: Implicit JOINs are implicit LEFT JOINs that are derived from to-one relationship path expressions. In order to e.g. access the COUNTRY columns from a CUSTOMER record, it is possible to write CUSTOMER.address().city().country().NAME. jOOQ will produce the necessary LEFT JOIN graph, which is much more tedious to write.
  • MULTISET_AGG function: The MULTISET aggregate function is not natively supported anywhere, even if a native implementation for the MULTISET value constructor exists.
  • Relational Division: Relational algebra supports a divison operator, which is the inverse operator of the cross product.
  • SEEK clause: The SEEK clause is a synthetic clause of the SELECT statement, which provides an alternative way of paginating other than the OFFSET clause. From a performance perspective, it is generally the preferred way to paginate.
  • SEMI JOIN and ANTI JOIN: Relational algebra defines SEMI JOIN and ANTI JOIN operators, which do not have a representation in any SQL dialect supported by jOOQ (Apache Impala has it, though). In SQL, the EXISTS predicate or IN predicate is used instead.
  • Sort indirection: When sorting, sometimes, we want to sort by a derived value, not the actual value of a column. Sort indirection makes this very easy with jOOQ.
  • UNIQUE predicate: This SQL standard predicate has not yet been implemented in any SQL dialect (it is being considered for H2). An esoteric, yet occasionally useful predicate that is difficult to emulate manually using the EXISTS predicate.
  • inline derived tables: A quick and type safe way to add a WHERE clause to a table expression for simplified dynamic SQL.
  • ON CONFLICT .. SET ALL TO EXCLUDED and ON DUPLICATE KEY .. SET ALL TO EXCLUDED: Convenience methods for the most common type of UPSERT statement.

Feedback

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

The jOOQ Logo