hcq/oceanbase
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

fix for hash groupby hung, exteng hashtable in process_popular_value_batch

Browse Source
This commit is contained in:
lucky-sinx
2024-11-05 08:17:24 +00:00
committed by ob-robot
parent 3768cfc728
commit 98d1dbb5a6
3 changed files with 16 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
src/sql/engine/aggregate/ob_exec_hash_struct_vec.cpp
View File

@ -254,7 +254,17 @@ int ObExtendHashTableVec<GroupRowBucket>::process_popular_value_batch(ObBatchRow
ret = OB_ERR_UNEXPECTED;
LOG_WARN("unexpected check_valid_threshold value", K(check_valid_threshold));
}
// extend bucket to hold whole batch
while (OB_SUCC(ret) && auto_extend_ && OB_UNLIKELY((size_ + result_brs->size_)
* SIZE_BUCKET_SCALE >= get_bucket_num())) {
int64_t pre_bkt_num = get_bucket_num();
if (OB_FAIL(extend())) {
SQL_ENG_LOG(WARN, "extend failed", K(ret));
} else if (get_bucket_num() <= pre_bkt_num) {
ret = OB_ERR_UNEXPECTED;
LOG_WARN("failed to extend table", K(ret), K(pre_bkt_num), K(get_bucket_num()));
}
}
for (int64_t i = 0; i < exprs.count(); ++i) {
if (nullptr == exprs.at(i)) {
vector_ptrs_.at(i) = nullptr;
@ -264,6 +274,9 @@ int ObExtendHashTableVec<GroupRowBucket>::process_popular_value_batch(ObBatchRow
}
for (int64_t i=0; OB_SUCC(ret) && i < result_brs->size_; i++) {
if (result_brs->skip_->at(i)) {
continue;
}
by_pass_rows++;
mask_hash = (hash_vals[i] & ObGroupRowBucketBase::HASH_VAL_MASK);
if (popular_map->size() == 0) {

4
src/sql/engine/aggregate/ob_hash_groupby_vec_op.cpp
View File

@ -2330,7 +2330,7 @@ int ObHashGroupByVecOp::by_pass_prepare_one_batch(const int64_t batch_size)
LOG_WARN("prepare add one row failed", K(ret));
}
for (int64_t i = 0; OB_SUCC(ret) && i < brs_.size_; i++) {
if (!batch_old_rows_[i]) {
if (!batch_old_rows_[i] || brs_.skip_->at(i)) {
continue;
}
by_pass_agg_rows_++;
@ -2509,7 +2509,7 @@ int ObHashGroupByVecOp::by_pass_restart_round()
if (skew_detection_enabled_ && bypass_ctrl_.by_passing()) {
// for popular values, so do not need to be big
OZ(local_group_rows_.resize(&mem_context_->get_malloc_allocator(),
INIT_BUCKET_COUNT_FOR_POPULAR));
INIT_BKT_SIZE_FOR_ADAPTIVE_GBY));
} else {
// if last round is in L2 cache, reuse the bucket
// otherwise resize to init size to avoid L2 cache overflow

1
src/sql/engine/aggregate/ob_hash_groupby_vec_op.h
View File

@ -126,7 +126,6 @@ public:
static const uint64_t MIN_CHECK_POPULAR_VALID_ROWS = 10000;
constexpr static const float SKEW_POPULAR_MAX_RATIO = 0.5;
const static int64_t SKEW_HEAP_SIZE = 15;
const static int64_t INIT_BUCKET_COUNT_FOR_POPULAR = 32;
public:
ObHashGroupByVecOp(ObExecContext &exec_ctx, const ObOpSpec &spec, ObOpInput *input)