Skip to content

[refactor](predicate) Disable predicates push down for cast expr #58960

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
Gabriel39 merged 2 commits into from
Dec 12, 2025
Merged

[refactor](predicate) Disable predicates push down for cast expr #58960

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c2bf6fe
[refactor](predicate) Disable predicates push down for cast expr
Gabriel39 Dec 11, 2025
edaf984
update
Gabriel39 Dec 11, 2025
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
49 changes: 7 additions & 42 deletions be/src/pipeline/exec/scan_operator.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -302,32 +302,28 @@ Status ScanLocalState<Derived>::_normalize_predicate(vectorized::VExprContext* c
static constexpr auto is_leaf = [](auto&& expr) { return !expr->is_and_expr(); };
auto in_predicate_checker = [&](const vectorized::VExprSPtrs& children,
SlotDescriptor** slot_desc, ColumnValueRangeType** range) {
if (children.empty() || vectorized::VExpr::expr_without_cast(children[0])->node_type() !=
TExprNodeType::SLOT_REF) {
if (children.empty() || children[0]->node_type() != TExprNodeType::SLOT_REF) {
// not a slot ref(column)
return false;
}
std::shared_ptr<vectorized::VSlotRef> slot =
std::dynamic_pointer_cast<vectorized::VSlotRef>(
vectorized::VExpr::expr_without_cast(children[0]));
std::dynamic_pointer_cast<vectorized::VSlotRef>(children[0]);
*slot_desc =
_parent->cast<typename Derived::Parent>()._slot_id_to_slot_desc[slot->slot_id()];
return _is_predicate_acting_on_slot(slot, children[0], range);
return _is_predicate_acting_on_slot(slot, range);
};
auto eq_predicate_checker = [&](const vectorized::VExprSPtrs& children,
SlotDescriptor** slot_desc, ColumnValueRangeType** range) {
if (children.empty() || vectorized::VExpr::expr_without_cast(children[0])->node_type() !=
TExprNodeType::SLOT_REF) {
if (children.empty() || children[0]->node_type() != TExprNodeType::SLOT_REF) {
// not a slot ref(column)
return false;
}
std::shared_ptr<vectorized::VSlotRef> slot =
std::dynamic_pointer_cast<vectorized::VSlotRef>(
vectorized::VExpr::expr_without_cast(children[0]));
std::dynamic_pointer_cast<vectorized::VSlotRef>(children[0]);
CHECK(slot != nullptr);
*slot_desc =
_parent->cast<typename Derived::Parent>()._slot_id_to_slot_desc[slot->slot_id()];
return _is_predicate_acting_on_slot(slot, children[0], range);
return _is_predicate_acting_on_slot(slot, range);
};

if (expr_root != nullptr) {
Expand Down Expand Up @@ -528,39 +524,19 @@ Status ScanLocalState<Derived>::_normalize_function_filters(vectorized::VExprCon

template <typename Derived>
bool ScanLocalState<Derived>::_is_predicate_acting_on_slot(
const std::shared_ptr<vectorized::VSlotRef>& slot_ref,
const vectorized::VExprSPtr& child_contains_slot, ColumnValueRangeType** range) {
const std::shared_ptr<vectorized::VSlotRef>& slot_ref, ColumnValueRangeType** range) {
auto entry = _slot_id_to_predicates.find(slot_ref->slot_id());
if (_slot_id_to_predicates.end() == entry) {
return false;
}
if (is_complex_type(slot_ref->data_type()->get_primitive_type())) {
return false;
}
auto& p = _parent->cast<typename Derived::Parent>();
auto sid_to_range = _slot_id_to_value_range.find(slot_ref->slot_id());
if (_slot_id_to_value_range.end() == sid_to_range) {
return false;
}
*range = &(sid_to_range->second);
SlotDescriptor* src_slot_desc = p._slot_id_to_slot_desc[slot_ref->slot_id()];
DCHECK(child_contains_slot != nullptr);
if (child_contains_slot->data_type()->get_primitive_type() !=
src_slot_desc->type()->get_primitive_type() ||
child_contains_slot->data_type()->get_precision() !=
src_slot_desc->type()->get_precision() ||
child_contains_slot->data_type()->get_scale() != src_slot_desc->type()->get_scale()) {
return _ignore_cast(src_slot_desc, child_contains_slot.get());
}
if ((child_contains_slot->data_type()->get_primitive_type() == PrimitiveType::TYPE_DATETIME ||
child_contains_slot->data_type()->get_primitive_type() == PrimitiveType::TYPE_DATETIMEV2 ||
child_contains_slot->data_type()->get_primitive_type() ==
PrimitiveType::TYPE_TIMESTAMPTZ) &&
child_contains_slot->node_type() == doris::TExprNodeType::CAST_EXPR) {
// Expr `CAST(CAST(datetime_col AS DATE) AS DATETIME) = datetime_literal` should not be
// push down.
return false;
}
return true;
}

Expand All @@ -581,17 +557,6 @@ std::string ScanLocalState<Derived>::debug_string(int indentation_level) const {
return fmt::to_string(debug_string_buffer);
}

template <typename Derived>
bool ScanLocalState<Derived>::_ignore_cast(SlotDescriptor* slot, vectorized::VExpr* expr) {
// only one level cast expr could push down for variant type
// check if expr is cast and it's children is slot
if (slot->type()->get_primitive_type() == PrimitiveType::TYPE_VARIANT) {
return expr->node_type() == TExprNodeType::CAST_EXPR &&
expr->children().at(0)->is_slot_ref();
}
return false;
}

template <typename Derived>
Status ScanLocalState<Derived>::_eval_const_conjuncts(vectorized::VExprContext* expr_ctx,
PushDownType* pdt) {
Expand Down
3 changes: 0 additions & 3 deletions be/src/pipeline/exec/scan_operator.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -263,7 +263,6 @@ class ScanLocalState : public ScanLocalStateBase {
PushDownType* pdt);

bool _is_predicate_acting_on_slot(const std::shared_ptr<vectorized::VSlotRef>& slot_ref,
const vectorized::VExprSPtr& child_contains_slot,
ColumnValueRangeType** range);

template <PrimitiveType T>
Expand All @@ -286,8 +285,6 @@ class ScanLocalState : public ScanLocalStateBase {
std::vector<std::shared_ptr<ColumnPredicate>>& predicates,
ColumnValueRange<T>& range, PushDownType* pdt);

bool _ignore_cast(SlotDescriptor* slot, vectorized::VExpr* expr);

template <bool IsFixed, PrimitiveType PrimitiveType, typename ChangeFixedValueRangeFunc>
Status _change_value_range(ColumnValueRange<PrimitiveType>& range, const void* value,
const ChangeFixedValueRangeFunc& func, const std::string& fn_name,
Expand Down
Loading