본문으로 바로가기
본문으로 바로가기

베타 및 실험적 기능

ClickHouse는 오픈 소스 프로젝트이므로 ClickHouse 직원뿐만 아니라 커뮤니티로부터도 많은 기여를 받습니다. 이러한 기여는 종종 서로 다른 속도로 개발되며, 일부 기능은 일반 제공(Generally Available, GA) 상태로 간주되기까지 긴 프로토타입 단계나 충분한 커뮤니티 피드백 및 반복 과정이 필요할 수 있습니다.

기능이 언제 일반 제공 상태로 분류될지 확실하지 않기 때문에 기능을 베타(Beta)실험적(Experimental) 두 가지 범주로 구분합니다.

베타(Beta) 기능은 ClickHouse 팀에서 공식적으로 지원합니다. 실험적(Experimental) 기능은 ClickHouse 팀 또는 커뮤니티 주도로 개발되는 초기 프로토타입이며, 공식적으로는 지원되지 않습니다.

아래 섹션에서는 베타(Beta)실험적(Experimental) 기능의 특성을 설명합니다.

베타 기능

  • 일반 제공(GA)을 목표로 활발히 개발 중입니다
  • 주요 알려진 이슈는 GitHub에서 추적할 수 있습니다
  • 기능은 향후 변경될 수 있습니다
  • ClickHouse Cloud에서 활성화되어 있을 수 있습니다
  • ClickHouse 팀에서 베타 기능을 지원합니다

아래에는 ClickHouse Cloud에서 베타로 분류되며, 사용 중인 ClickHouse Cloud 서비스에서 사용할 수 있는 기능들이 정리되어 있습니다.

참고: 최근에 도입된 기능을 사용하려면 ClickHouse의 최신 버전 compatibility 설정을 사용하고 있는지 확인하십시오.

실험적 기능

  • GA(General Availability) 단계에 도달하지 못할 수도 있음
  • 제거될 수 있음
  • 호환성을 깨는 변경 사항을 일으킬 수 있음
  • 향후 동작이 변경될 수 있음
  • 명시적으로 활성화해야 함
  • ClickHouse 팀은 실험적 기능을 지원하지 않습니다
  • 중요한 기능이나 문서가 부족할 수 있음
  • 클라우드 환경에서는 활성화할 수 없음

유의하십시오: 위에서 Beta로 나열된 항목 이외의 추가 실험적 기능은 ClickHouse Cloud에서 활성화할 수 없습니다.

베타 설정

이름기본값
geotoh3_argument_orderlat_lon
enable_lightweight_update1
allow_experimental_correlated_subqueries1
parallel_replicas_count0
parallel_replica_offset0
parallel_replicas_custom_key``
parallel_replicas_custom_key_range_lower0
parallel_replicas_custom_key_range_upper0
parallel_replicas_filter_pushdown0
allow_experimental_database_iceberg0
allow_experimental_database_unity_catalog0
allow_experimental_database_glue_catalog0
session_timezone``
low_priority_query_wait_time_ms1000
allow_statistics_optimize1
use_statistics1
use_statistics_cache1
enable_full_text_index1
allow_experimental_delta_kernel_rs1
allow_insert_into_iceberg0
enable_join_runtime_filters1

실험적 설정

이름기본값
allow_experimental_replacing_merge_with_cleanup0
allow_experimental_reverse_key0
allow_remote_fs_zero_copy_replication0
distributed_index_analysis_min_indexes_bytes_to_activate1073741824
distributed_index_analysis_min_parts_to_activate10
enable_replacing_merge_with_cleanup_for_min_age_to_force_merge0
force_read_through_cache_for_merges0
merge_selector_algorithmSimple
merge_selector_enable_heuristic_to_lower_max_parts_to_merge_at_once0
merge_selector_heuristic_to_lower_max_parts_to_merge_at_once_exponent5
notify_newest_block_number0
part_moves_between_shards_delay_seconds30
part_moves_between_shards_enable0
remote_fs_zero_copy_path_compatible_mode0
remote_fs_zero_copy_zookeeper_path/clickhouse/zero_copy
remove_rolled_back_parts_immediately1
shared_merge_tree_virtual_parts_discovery_batch1
allow_experimental_kafka_offsets_storage_in_keeper0
allow_experimental_delta_lake_writes0
automatic_parallel_replicas_mode0
automatic_parallel_replicas_min_bytes_per_replica1048576
distributed_index_analysis0
allow_experimental_materialized_postgresql_table0
allow_experimental_funnel_functions0
allow_experimental_nlp_functions0
allow_experimental_hash_functions0
allow_experimental_time_series_table0
allow_experimental_codecs0
throw_on_unsupported_query_inside_transaction1
wait_changes_become_visible_after_commit_modewait_unknown
implicit_transaction0
grace_hash_join_initial_buckets1
grace_hash_join_max_buckets1024
join_to_sort_minimum_perkey_rows40
join_to_sort_maximum_table_rows10000
allow_experimental_join_right_table_sorting0
allow_experimental_statistics0
allow_experimental_window_view0
window_view_clean_interval60
window_view_heartbeat_interval15
wait_for_window_view_fire_signal_timeout10
stop_refreshable_materialized_views_on_startup0
allow_experimental_database_materialized_postgresql0
allow_experimental_nullable_tuple_type0
allow_experimental_query_deduplication0
allow_experimental_database_hms_catalog0
allow_experimental_kusto_dialect0
allow_experimental_prql_dialect0
enable_adaptive_memory_spill_scheduler0
allow_experimental_iceberg_compaction0
write_full_path_in_iceberg_metadata0
iceberg_metadata_compression_method``
make_distributed_plan0
distributed_plan_execute_locally0
distributed_plan_default_shuffle_join_bucket_count8
distributed_plan_default_reader_bucket_count8
distributed_plan_force_exchange_kind``
distributed_plan_max_rows_to_broadcast20000
allow_experimental_ytsaurus_table_engine0
allow_experimental_ytsaurus_table_function0
allow_experimental_ytsaurus_dictionary_source0
distributed_plan_force_shuffle_aggregation0
join_runtime_filter_exact_values_limit10000
join_runtime_bloom_filter_bytes524288
join_runtime_bloom_filter_hash_functions3
join_runtime_filter_pass_ratio_threshold_for_disabling0.7
join_runtime_filter_blocks_to_skip_before_reenabling30
join_runtime_bloom_filter_max_ratio_of_set_bits0.7
rewrite_in_to_join0
allow_experimental_time_series_aggregate_functions0
promql_database``
promql_table``
promql_evaluation_timeauto
allow_experimental_alias_table_engine0
use_paimon_partition_pruning0
allow_experimental_object_storage_queue_hive_partitioning0
query_plan_optimize_join_order_algorithmgreedy
allow_experimental_database_paimon_rest_catalog0