commit 901a0bf733d26761bd47f2034bf62d86e3fb3f2c
parent 6a7a49d6fa0a03f61c14f34ff0c34911dbd9ba09
Author: a3nm <>
Date: Sun, 17 Jun 2018 02:20:07 +0200
non-boolean, etc.
Diffstat:
1 file changed, 18 insertions(+), 0 deletions(-)
diff --git a/Query evaluation.page b/Query evaluation.page
@@ -0,0 +1,17 @@
+The task of **query evaluation** is concerned with the complexity of determining whether a database satisfies a logical formula. In other words, it is the analogue of the logical problem of [model checking]().
+
+# Complexity measures
+
+TODO: combined complexity, data complexity, query complexity
+
+# Basic results
+
+TODO
+
+# Non-Boolean queries
+
+It is often most convenient to assume that queries are Boolean, so query evaluation returns a YES/NO answer. If a query $Q(\mathbf{x})$ is non-Boolean, we can always reduce to the Boolean case by considering all possible results of the query, i.e., all tuples $\mathbf{a}$ of the active domain of the database instance, and solving query evaluation for each of the Boolean queries $Q(\mathbf{a})$. Note that this reduction is in [PTIME](Complexity classes) if the arity of the query is fixed.
+
+A different perspective on non-Boolean queries is to write the query in [relational algebra]() and consider the evaluation of relational algebra instead.
+
+Yet another perspective is to study the complexity of producing the query results in succession: see [Enumeration algorithms]().+
\ No newline at end of file
