Predicates are currently created as ASTs, but are not compiled or optimized. Instead, evaluating them means traversing the tree in its entirety, something that, in a tight loop (like going over partitioning functions for all files in a table), is suboptimal.
Compiling the expressions into a minimal syntax tree that embeds the right-hand side values would let us optimize it ahead of time in several ways, including:
Compressing equals-or-equals trees to in.
Dropping always-true and always-false expressions entirely.
This issue therefore includes three things:
Representing expressions as a value-embedded AST.
Creating an optimizer framework.
Slowly building more and more optimizations.
This could probably be done using existing tools, without reinventing the wheel :)
Predicates are currently created as ASTs, but are not compiled or optimized. Instead, evaluating them means traversing the tree in its entirety, something that, in a tight loop (like going over partitioning functions for all files in a table), is suboptimal.
Compiling the expressions into a minimal syntax tree that embeds the right-hand side values would let us optimize it ahead of time in several ways, including:
equals-or-equalstrees toin.trueand always-falseexpressions entirely.This issue therefore includes three things:
This could probably be done using existing tools, without reinventing the wheel :)