mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-11-28 02:21:59 +00:00
More accurate prediction. [#CLICKHOUSE-2902]
This commit is contained in:
parent
f519100946
commit
c78177a28a
@ -71,6 +71,9 @@ Block MergeTreeBaseBlockInputStream::readFromPart()
|
||||
size_t res_block_size_bytes = 0;
|
||||
bool bytes_exceeded = false;
|
||||
|
||||
if (task->size_predictor)
|
||||
task->size_predictor->startBlock();
|
||||
|
||||
if (prewhere_actions)
|
||||
{
|
||||
do
|
||||
|
@ -127,12 +127,6 @@ MergeTreeBlockSizePredictor::MergeTreeBlockSizePredictor(
|
||||
|
||||
void MergeTreeBlockSizePredictor::startBlock()
|
||||
{
|
||||
if (!is_initialized)
|
||||
{
|
||||
is_initialized = true;
|
||||
//init();
|
||||
}
|
||||
|
||||
block_size_bytes = 0;
|
||||
block_size_rows = 0;
|
||||
for (auto & info : dynamic_columns_infos)
|
||||
@ -147,16 +141,16 @@ void MergeTreeBlockSizePredictor::update(const Block& block, size_t read_marks)
|
||||
block_size_bytes = block_size_rows * fixed_columns_bytes_per_row;
|
||||
bytes_per_row_current = fixed_columns_bytes_per_row;
|
||||
|
||||
/// Make recursive updates for each read mark
|
||||
double alpha = std::pow(1. - decay, read_marks);
|
||||
|
||||
for (auto & info : dynamic_columns_infos)
|
||||
{
|
||||
size_t new_size = block.getByName(info.name).column->byteSize();
|
||||
size_t dif_size = new_size - info.size_bytes;
|
||||
|
||||
double local_bytes_per_row = static_cast<double>(dif_size) / dif_rows;
|
||||
|
||||
/// Make recursive updates for each read mark
|
||||
for (size_t i = 0; i < read_marks; ++i)
|
||||
info.bytes_per_row += decay * (local_bytes_per_row - info.bytes_per_row);
|
||||
info.bytes_per_row = alpha * info.bytes_per_row + (1. - alpha) * local_bytes_per_row;
|
||||
|
||||
info.size_bytes = new_size;
|
||||
block_size_bytes += new_size;
|
||||
|
@ -80,9 +80,11 @@ struct MergeTreeBlockSizePredictor
|
||||
}
|
||||
|
||||
MergeTreeData::DataPartPtr data_part;
|
||||
bool is_initialized = false;
|
||||
size_t index_granularity;
|
||||
const double decay = 0.75;
|
||||
|
||||
/// Aggressiveness of bytes_per_row updates
|
||||
/// One update per mark
|
||||
static constexpr double decay = 0.5;
|
||||
|
||||
struct ColumnInfo
|
||||
{
|
||||
|
Loading…
Reference in New Issue
Block a user