database/doris
2
0
Fork 0
Code Issues Pull Requests Actions 3 Packages Projects Releases Wiki Activity

[chore](key_util) remove useless null_first parameter (#26635)

Browse Source 
Doris always put null in the first when sorting key, the parameter null_first of encode_keys is useless.
This commit is contained in:
zhannngchen
2023-11-10 14:27:47 +08:00
committed by GitHub
parent cdba4936b4
commit 899630d0eb
5 changed files with 14 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
be/src/olap/rowset/segment_v2/segment_iterator.cpp
View File

@ -1184,7 +1184,7 @@ Status SegmentIterator::_lookup_ordinal_from_pk_index(const RowCursor& key, bool
std::string index_key;
// when is_include is false, we shoudle append KEY_NORMAL_MARKER to the
// encode key. Otherwise, we will get an incorrect upper bound.
encode_key_with_padding<RowCursor, true, true>(
encode_key_with_padding<RowCursor, true>(
&index_key, key, _segment->_tablet_schema->num_key_columns(), is_include, true);
if (index_key < _segment->min_key()) {
*rowid = 0;

20
be/src/olap/rowset/segment_v2/segment_writer.cpp
View File

@ -737,8 +737,7 @@ int64_t SegmentWriter::max_row_to_add(size_t row_avg_size_in_bytes) {
}
std::string SegmentWriter::_full_encode_keys(
const std::vector<vectorized::IOlapColumnDataAccessor*>& key_columns, size_t pos,
bool null_first) {
const std::vector<vectorized::IOlapColumnDataAccessor*>& key_columns, size_t pos) {
assert(_key_index_size.size() == _num_key_columns);
assert(key_columns.size() == _num_key_columns && _key_coders.size() == _num_key_columns);
@ -747,11 +746,7 @@ std::string SegmentWriter::_full_encode_keys(
for (const auto& column : key_columns) {
auto field = column->get_data_at(pos);
if (UNLIKELY(!field)) {
if (null_first) {
encoded_keys.push_back(KEY_NULL_FIRST_MARKER);
} else {
encoded_keys.push_back(KEY_NULL_LAST_MARKER);
}
encoded_keys.push_back(KEY_NULL_FIRST_MARKER);
++cid;
continue;
}
@ -779,8 +774,7 @@ void SegmentWriter::_encode_seq_column(const vectorized::IOlapColumnDataAccessor
}
std::string SegmentWriter::_encode_keys(
const std::vector<vectorized::IOlapColumnDataAccessor*>& key_columns, size_t pos,
bool null_first) {
const std::vector<vectorized::IOlapColumnDataAccessor*>& key_columns, size_t pos) {
assert(key_columns.size() == _num_short_key_columns);
std::string encoded_keys;
@ -788,11 +782,7 @@ std::string SegmentWriter::_encode_keys(
for (const auto& column : key_columns) {
auto field = column->get_data_at(pos);
if (UNLIKELY(!field)) {
if (null_first) {
encoded_keys.push_back(KEY_NULL_FIRST_MARKER);
} else {
encoded_keys.push_back(KEY_NULL_LAST_MARKER);
}
encoded_keys.push_back(KEY_NULL_FIRST_MARKER);
++cid;
continue;
}
@ -810,7 +800,7 @@ Status SegmentWriter::append_row(const RowType& row) {
RETURN_IF_ERROR(_column_writers[cid]->append(cell));
}
std::string full_encoded_key;
encode_key<RowType, true, true>(&full_encoded_key, row, _num_key_columns);
encode_key<RowType, true>(&full_encoded_key, row, _num_key_columns);
if (_tablet_schema->has_sequence_col()) {
full_encoded_key.push_back(KEY_NORMAL_MARKER);
auto cid = _tablet_schema->sequence_col_idx();

5
be/src/olap/rowset/segment_v2/segment_writer.h
View File

@ -146,11 +146,10 @@ private:
Status _write_raw_data(const std::vector<Slice>& slices);
void _maybe_invalid_row_cache(const std::string& key);
std::string _encode_keys(const std::vector<vectorized::IOlapColumnDataAccessor*>& key_columns,
size_t pos, bool null_first = true);
size_t pos);
// used for unique-key with merge on write and segment min_max key
std::string _full_encode_keys(
const std::vector<vectorized::IOlapColumnDataAccessor*>& key_columns, size_t pos,
bool null_first = true);
const std::vector<vectorized::IOlapColumnDataAccessor*>& key_columns, size_t pos);
// used for unique-key with merge on write
void _encode_seq_column(const vectorized::IOlapColumnDataAccessor* seq_column, size_t pos,
string* encoded_keys);

5
be/src/service/point_query_executor.cpp
View File

@ -254,9 +254,8 @@ Status PointQueryExecutor::_init_keys(const PTabletKeyLookupRequest* request) {
RowCursor cursor;
RETURN_IF_ERROR(cursor.init_scan_key(_tablet->tablet_schema(), olap_tuples[i].values()));
RETURN_IF_ERROR(cursor.from_tuple(olap_tuples[i]));
encode_key_with_padding<RowCursor, true, true>(&_row_read_ctxs[i]._primary_key, cursor,
_tablet->tablet_schema()->num_key_columns(),
true);
encode_key_with_padding<RowCursor, true>(&_row_read_ctxs[i]._primary_key, cursor,
_tablet->tablet_schema()->num_key_columns(), true);
}
return Status::OK();
}

18
be/src/util/key_util.h
View File

@ -48,8 +48,6 @@ constexpr uint8_t KEY_MINIMAL_MARKER = 0x00;
constexpr uint8_t KEY_NULL_FIRST_MARKER = 0x01;
// Used to represent a normal field, which content is encoded after this marker
constexpr uint8_t KEY_NORMAL_MARKER = 0x02;
// Used to represent
constexpr uint8_t KEY_NULL_LAST_MARKER = 0xFE;
// Used to represent maximal value for that field
constexpr uint8_t KEY_MAXIMAL_MARKER = 0xFF;
@ -61,7 +59,7 @@ constexpr uint8_t KEY_MAXIMAL_MARKER = 0xFF;
// If all num_keys are found in row, no marker will be added.
// if padding_minimal is false and padding_normal_marker is true,
// KEY_NORMAL_MARKER will be added.
template <typename RowType, bool null_first = true, bool full_encode = false>
template <typename RowType, bool full_encode = false>
void encode_key_with_padding(std::string* buf, const RowType& row, size_t num_keys,
bool padding_minimal, bool padding_normal_marker = false) {
for (auto cid = 0; cid < num_keys; cid++) {
@ -80,11 +78,7 @@ void encode_key_with_padding(std::string* buf, const RowType& row, size_t num_ke
auto cell = row.cell(cid);
if (cell.is_null()) {
if (null_first) {
buf->push_back(KEY_NULL_FIRST_MARKER);
} else {
buf->push_back(KEY_NULL_LAST_MARKER);
}
buf->push_back(KEY_NULL_FIRST_MARKER);
continue;
}
buf->push_back(KEY_NORMAL_MARKER);
@ -99,16 +93,12 @@ void encode_key_with_padding(std::string* buf, const RowType& row, size_t num_ke
// Encode one row into binary according given num_keys.
// Client call this function must assure that row contains the first
// num_keys columns.
template <typename RowType, bool null_first = true, bool full_encode = false>
template <typename RowType, bool full_encode = false>
void encode_key(std::string* buf, const RowType& row, size_t num_keys) {
for (auto cid = 0; cid < num_keys; cid++) {
auto cell = row.cell(cid);
if (cell.is_null()) {
if (null_first) {
buf->push_back(KEY_NULL_FIRST_MARKER);
} else {
buf->push_back(KEY_NULL_LAST_MARKER);
}
buf->push_back(KEY_NULL_FIRST_MARKER);
continue;
}
buf->push_back(KEY_NORMAL_MARKER);