Idempotence: Difference between revisions
imported>Howard C. Berkowitz (Very nice text! linked a bit; I'll need to expand on ACID properties.) |
Pat Palmer (talk | contribs) m (→In mathematics) |
||
Line 10: | Line 10: | ||
equivalently, every element is an [[idempotent element]] for <math>\star</math>. | equivalently, every element is an [[idempotent element]] for <math>\star</math>. | ||
Examples of idempotent binary operations include [[join]] and [[meet]] in a [[lattice (order)|lattice]]; [[union]] and [[intersection]] on [[set (mathematics)|sets]]; [[disjunction]] and [[conjunction]] in [[propositional logic]]. | Examples of idempotent binary operations include [[join]] and [[meet]] in a [[lattice (order)|lattice]]; [[union]] and [[intersection]] on [[set (mathematics)|sets]]; [[disjunction]] and [[Conjunction (logical and)|conjunction]] in [[propositional logic]]. | ||
A [[unary operation]] (a [[function (mathematics)|function]] from a set to itself) π is idempotent if it is an idempotent element for [[function composition]], <math>\pi \circ \pi = \pi</math>. | A [[unary operation]] (a [[function (mathematics)|function]] from a set to itself) π is idempotent if it is an idempotent element for [[function composition]], <math>\pi \circ \pi = \pi</math>. |
Revision as of 12:20, 18 November 2022
In mathematics and computer science idempotence is the property of an operation that repeated application has no further effect.
In mathematics
A binary operation is idempotent if
- for all x:
equivalently, every element is an idempotent element for .
Examples of idempotent binary operations include join and meet in a lattice; union and intersection on sets; disjunction and conjunction in propositional logic.
A unary operation (a function from a set to itself) π is idempotent if it is an idempotent element for function composition, .
In computing
In applications such as databases and transaction processing, idempotent operations are those for which the intended effect is that repeated application should have no effect, such as inserting a record into a file, an element into a set, or sending a message. Implementations must therefore be constructed in such a way that the intended effect is actually carried into practice. For example, messages might have unique sequence numbers with duplicates being discarded on receipt; a set might be implemented as a bit vector, and member insertion implemented by an idempotent mathematical operation such as inclusive or with a bit mask.
When a particular unit of work (i.e., transaction), has the idempotent property, relaxation of the ACID properties usually required for reliable transaction processing, can be relaxed.