This is an automated email from the ASF dual-hosted git repository.
kxiao pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/master by this push:
new cbcb81f3813 [FIX](complextype)fix compare_at base function support
nested types (#29297)
cbcb81f3813 is described below
commit cbcb81f3813934e72f4fe949f3d026429a216821
Author: amory <amorywang...@gmail.com>
AuthorDate: Sat Jan 6 12:05:43 2024 +0800
[FIX](complextype)fix compare_at base function support nested types (#29297)
---
be/src/vec/columns/column.h | 1 +
be/src/vec/columns/column_array.cpp | 20 ++++++++++++++++++++
be/src/vec/columns/column_array.h | 7 ++-----
be/src/vec/columns/column_map.cpp | 29 +++++++++++++++++++++++++++++
be/src/vec/columns/column_map.h | 7 ++-----
be/src/vec/columns/column_struct.cpp | 15 +++++++++++++++
be/src/vec/columns/column_struct.h | 6 +-----
7 files changed, 70 insertions(+), 15 deletions(-)
diff --git a/be/src/vec/columns/column.h b/be/src/vec/columns/column.h
index 25ad1145fc8..d67ad3e206b 100644
--- a/be/src/vec/columns/column.h
+++ b/be/src/vec/columns/column.h
@@ -453,6 +453,7 @@ public:
* For example, if nan_direction_hint == -1 is used by descending
sorting, NaNs will be at the end.
*
* For non Nullable and non floating point types, nan_direction_hint is
ignored.
+ * For array/map/struct types, we compare with nested column element and
offsets size
*/
virtual int compare_at(size_t n, size_t m, const IColumn& rhs,
int nan_direction_hint) const = 0;
diff --git a/be/src/vec/columns/column_array.cpp
b/be/src/vec/columns/column_array.cpp
index 0c926c0f31e..866a55c447f 100644
--- a/be/src/vec/columns/column_array.cpp
+++ b/be/src/vec/columns/column_array.cpp
@@ -254,6 +254,26 @@ StringRef ColumnArray::serialize_value_into_arena(size_t
n, Arena& arena,
return res;
}
+int ColumnArray::compare_at(size_t n, size_t m, const IColumn& rhs_, int
nan_direction_hint) const {
+ // since column type is complex, we can't use this function
+ const auto& rhs = assert_cast<const ColumnArray&>(rhs_);
+
+ size_t lhs_size = size_at(n);
+ size_t rhs_size = rhs.size_at(m);
+ size_t min_size = std::min(lhs_size, rhs_size);
+ for (size_t i = 0; i < min_size; ++i) {
+ if (int res = get_data().compare_at(offset_at(n) + i, rhs.offset_at(m)
+ i, *rhs.data.get(),
+ nan_direction_hint);
+ res) {
+ // if res != 0 , here is something different ,just return
+ return res;
+ }
+ }
+
+ // then we check size of array
+ return lhs_size < rhs_size ? -1 : (lhs_size == rhs_size ? 0 : 1);
+}
+
const char* ColumnArray::deserialize_and_insert_from_arena(const char* pos) {
size_t array_size = unaligned_load<size_t>(pos);
pos += sizeof(array_size);
diff --git a/be/src/vec/columns/column_array.h
b/be/src/vec/columns/column_array.h
index ad8edbf7565..f2f187a7236 100644
--- a/be/src/vec/columns/column_array.h
+++ b/be/src/vec/columns/column_array.h
@@ -163,11 +163,8 @@ public:
//ColumnPtr index(const IColumn & indexes, size_t limit) const;
template <typename Type>
ColumnPtr index_impl(const PaddedPODArray<Type>& indexes, size_t limit)
const;
- [[noreturn]] int compare_at(size_t n, size_t m, const IColumn& rhs_,
- int nan_direction_hint) const override {
- LOG(FATAL) << "compare_at not implemented";
- __builtin_unreachable();
- }
+ int compare_at(size_t n, size_t m, const IColumn& rhs_, int
nan_direction_hint) const override;
+
[[noreturn]] void get_permutation(bool reverse, size_t limit, int
nan_direction_hint,
Permutation& res) const override {
LOG(FATAL) << "get_permutation not implemented";
diff --git a/be/src/vec/columns/column_map.cpp
b/be/src/vec/columns/column_map.cpp
index 6981c25fc23..e46ea7fe683 100644
--- a/be/src/vec/columns/column_map.cpp
+++ b/be/src/vec/columns/column_map.cpp
@@ -231,6 +231,35 @@ const char*
ColumnMap::deserialize_and_insert_from_arena(const char* pos) {
return pos;
}
+int ColumnMap::compare_at(size_t n, size_t m, const IColumn& rhs_, int
nan_direction_hint) const {
+ const auto& rhs = assert_cast<const ColumnMap&>(rhs_);
+
+ size_t lhs_size = size_at(n);
+ size_t rhs_size = rhs.size_at(m);
+
+ size_t lhs_offset = offset_at(n);
+ size_t rhs_offset = rhs.offset_at(m);
+
+ size_t min_size = std::min(lhs_size, rhs_size);
+
+ for (size_t i = 0; i < min_size; ++i) {
+ // if any value in key not equal, just return
+ if (int res = get_keys().compare_at(lhs_offset + i, rhs_offset + i,
rhs.get_keys(),
+ nan_direction_hint);
+ res) {
+ return res;
+ }
+ // // if any value in value not equal, just return
+ if (int res = get_values().compare_at(lhs_offset + i, rhs_offset + i,
rhs.get_values(),
+ nan_direction_hint);
+ res) {
+ return res;
+ }
+ }
+
+ return lhs_size < rhs_size ? -1 : (lhs_size == rhs_size ? 0 : 1);
+}
+
void ColumnMap::update_hash_with_value(size_t n, SipHash& hash) const {
size_t kv_size = size_at(n);
size_t offset = offset_at(n);
diff --git a/be/src/vec/columns/column_map.h b/be/src/vec/columns/column_map.h
index 59725111d73..45cedeb0d94 100644
--- a/be/src/vec/columns/column_map.h
+++ b/be/src/vec/columns/column_map.h
@@ -119,11 +119,8 @@ public:
return scatter_impl<ColumnMap>(num_columns, selector);
}
- [[noreturn]] int compare_at(size_t n, size_t m, const IColumn& rhs_,
- int nan_direction_hint) const override {
- LOG(FATAL) << "compare_at not implemented";
- __builtin_unreachable();
- }
+ int compare_at(size_t n, size_t m, const IColumn& rhs_, int
nan_direction_hint) const override;
+
void get_permutation(bool reverse, size_t limit, int nan_direction_hint,
Permutation& res) const override {
LOG(FATAL) << "get_permutation not implemented";
diff --git a/be/src/vec/columns/column_struct.cpp
b/be/src/vec/columns/column_struct.cpp
index 5a89b5d754c..4e69db0a545 100644
--- a/be/src/vec/columns/column_struct.cpp
+++ b/be/src/vec/columns/column_struct.cpp
@@ -185,6 +185,21 @@ const char*
ColumnStruct::deserialize_and_insert_from_arena(const char* pos) {
return pos;
}
+int ColumnStruct::compare_at(size_t n, size_t m, const IColumn& rhs_,
+ int nan_direction_hint) const {
+ const ColumnStruct& rhs = assert_cast<const ColumnStruct&>(rhs_);
+
+ const size_t lhs_tuple_size = columns.size();
+ const size_t rhs_tuple_size = rhs.tuple_size();
+ const size_t min_size = std::min(lhs_tuple_size, rhs_tuple_size);
+ for (size_t i = 0; i < min_size; ++i) {
+ if (int res = columns[i]->compare_at(n, m, *rhs.columns[i],
nan_direction_hint); res) {
+ return res;
+ }
+ }
+ return lhs_tuple_size > rhs_tuple_size ? 1 : (lhs_tuple_size <
rhs_tuple_size ? -1 : 0);
+}
+
void ColumnStruct::update_hash_with_value(size_t n, SipHash& hash) const {
for (const auto& column : columns) {
column->update_hash_with_value(n, hash);
diff --git a/be/src/vec/columns/column_struct.h
b/be/src/vec/columns/column_struct.h
index a3141521e50..73f40713729 100644
--- a/be/src/vec/columns/column_struct.h
+++ b/be/src/vec/columns/column_struct.h
@@ -158,11 +158,7 @@ public:
MutableColumns scatter(ColumnIndex num_columns, const Selector& selector)
const override;
// ColumnPtr index(const IColumn & indexes, size_t limit) const override;
- [[noreturn]] int compare_at(size_t n, size_t m, const IColumn& rhs_,
- int nan_direction_hint) const override {
- LOG(FATAL) << "compare_at not implemented";
- __builtin_unreachable();
- }
+ int compare_at(size_t n, size_t m, const IColumn& rhs_, int
nan_direction_hint) const override;
MutableColumnPtr get_shrinked_column() override;
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
