Skip to content

[opt](in expr) Optimize the IN expression by skipping constant column… #40917

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
HappenLee merged 1 commit into from
Sep 18, 2024
Merged

[opt](in expr) Optimize the IN expression by skipping constant column… #40917

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 16 additions & 3 deletions be/src/vec/exprs/vin_predicate.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -93,10 +93,24 @@ Status VInPredicate::open(RuntimeState* state, VExprContext* context,
if (scope == FunctionContext::FRAGMENT_LOCAL) {
RETURN_IF_ERROR(VExpr::get_const_col(context, nullptr));
}

_is_args_all_constant = std::all_of(_children.begin() + 1, _children.end(),
[](const VExprSPtr& expr) { return expr->is_constant(); });
_open_finished = true;
return Status::OK();
}

size_t VInPredicate::skip_constant_args_size() const {
if (_is_args_all_constant && !_can_fast_execute) {
// This is an optimization. For expressions like colA IN (1, 2, 3, 4),
// where all values inside the IN clause are constants,
// a hash set is created during open, and it will not be accessed again during execute
// Here, _children[0] is colA
return 1;
}
return _children.size();
}

void VInPredicate::close(VExprContext* context, FunctionContext::FunctionStateScope scope) {
VExpr::close_function_context(context, scope, _function);
VExpr::close(context, scope);
Expand All @@ -115,9 +129,8 @@ Status VInPredicate::execute(VExprContext* context, Block* block, int* result_co
return Status::OK();
}
DCHECK(_open_finished || _getting_const_col);
// TODO: not execute const expr again, but use the const column in function context
doris::vectorized::ColumnNumbers arguments(_children.size());
for (int i = 0; i < _children.size(); ++i) {
doris::vectorized::ColumnNumbers arguments(skip_constant_args_size());
for (int i = 0; i < skip_constant_args_size(); ++i) {
int column_id = -1;
RETURN_IF_ERROR(_children[i]->execute(context, block, &column_id));
arguments[i] = column_id;
Expand Down
3 changes: 3 additions & 0 deletions be/src/vec/exprs/vin_predicate.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,6 +51,8 @@ class VInPredicate final : public VExpr {

std::string debug_string() const override;

size_t skip_constant_args_size() const;

const FunctionBasePtr function() { return _function; }

bool is_not_in() const { return _is_not_in; };
Expand All @@ -62,5 +64,6 @@ class VInPredicate final : public VExpr {

const bool _is_not_in;
static const constexpr char* function_name = "in";
bool _is_args_all_constant = false;
};
} // namespace doris::vectorized