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
IN-list Padding
Applies to ✅ Open Source Edition ✅ Express Edition ✅ Professional Edition ✅ Enterprise Edition
Databases that feature a cursor cache / statement cache (e.g. Oracle, SQL Server, DB2, etc.) are highly optimised for prepared statement re-use. When a client sends a prepared statement to the server, the server will go to the cache and look up whether there already exists a previously calculated execution plan for the statement (i.e. the SQL string). This is called a "soft-parse" (in Oracle). If not, the execution plan is calculated on the fly. This is called a "hard-parse" (in Oracle).
Preventing hard-parses is extremely important in high throughput OLTP systems where queries are usually not very complex but are run millions of times in a short amount of time. Using bind variables, this is usually not a problem, with the exception of the IN predicate, which generates different SQL strings even when using bind variables:
-- All of these are different SQL statements: SELECT * FROM AUTHOR WHERE ID IN (?) SELECT * FROM AUTHOR WHERE ID IN (?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?, ?)
This problem may not be obvious to Java / jOOQ developers, as they are always produced from the same jOOQ statement:
// All of these are the same jOOQ statement DSL.using(configuration) .select() .from(AUTHOR) .where(AUTHOR.ID.in(collection)) .fetch();
Depending on the possible sizes of the collection, it may be worth exploring using arrays or temporary tables as a workaround, or to reuse the original query that produced the set of IDs in the first place (through a semi-join). But sometimes, this is not possible. In this case, users can opt in to a third workaround: enabling the inListPadding
setting. If enabled, jOOQ will "pad" the IN
list to a length that is a power of two (configurable with Settings.inListPadBase
). So, the original queries would look like this instead:
-- Original SELECT * FROM AUTHOR WHERE ID IN (?) SELECT * FROM AUTHOR WHERE ID IN (?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?, ?, ?)
-- Padded SELECT * FROM AUTHOR WHERE ID IN (?) SELECT * FROM AUTHOR WHERE ID IN (?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?, ?, ?, ?, ?) SELECT * FROM AUTHOR WHERE ID IN (?, ?, ?, ?, ?, ?, ?, ?)
This technique will drastically reduce the number of possible SQL strings without impairing too much the usual cases where the IN
list is small. When padding, the last bind variable will simply be repeated many times.
Usually, there is a better way - use this as a last resort!
Example configuration
Settings settings = new Settings() .withInListPadding(true) // Default to false .withInListPadBase(4); // Default to 2
Feedback
Do you have any feedback about this page? We'd love to hear it!