测试版和实验功能

因为 ClickHouse 是开源的，所以它不仅得到了 ClickHouse 员工的贡献，也得到了社区的贡献。这些贡献通常以不同的速度进行；某些功能可能需要较长的原型阶段或更多时间以获得足够的社区反馈和迭代，才能被视为正式可用（GA）。

由于何时将功能归类为正式可用存在不确定性，我们将功能划分为两个类别：Beta 和 Experimental。

Beta 功能得到了 ClickHouse 团队的正式支持。 Experimental 功能是由 ClickHouse 团队或社区驱动的早期原型，并未得到正式支持。

以下部分明确描述了 Beta 和 Experimental 功能的属性：

Beta Features

• 正在积极开发中，以使其正式可用（GA）
• 主要已知问题可以在 GitHub 上跟踪
• 功能可能在未来发生变化
• 可能在 ClickHouse Cloud 中启用
• ClickHouse 团队支持 beta 功能

以下功能被视为 ClickHouse Cloud 中的 Beta，并可在 ClickHouse Cloud 服务中使用，即使它们可能当前处于 ClickHouse 的 allow_experimental_* 设置下：

注意：请确保您使用的是当前版本的 ClickHouse compatibility 设置，以便使用最近引入的功能。

Experimental Features

• 可能永远不会成为 GA
• 可能会被移除
• 可以引入重大更改
• 功能可能在未来发生变化
• 需要明确启用
• ClickHouse 团队不支持 experimental 功能
• 可能缺少重要的功能和文档
• 不能在云中启用

请注意：除了上述列为 Beta 的功能外，ClickHouse Cloud 中不允许启用其他 experimental 功能。

Beta settings

Name	Default
allow_remote_fs_zero_copy_replication	0
geotoh3_argument_order	lat_lon
allow_experimental_parallel_reading_from_replicas	0
parallel_replicas_mode	read_tasks
parallel_replicas_count	0
parallel_replica_offset	0
parallel_replicas_custom_key	``
parallel_replicas_custom_key_range_lower	0
parallel_replicas_custom_key_range_upper	0
cluster_for_parallel_replicas	``
parallel_replicas_allow_in_with_subquery	1
parallel_replicas_for_non_replicated_merge_tree	0
parallel_replicas_min_number_of_rows_per_replica	0
parallel_replicas_prefer_local_join	1
parallel_replicas_mark_segment_size	0
parallel_replicas_local_plan	1
parallel_replicas_index_analysis_only_on_coordinator	1
parallel_replicas_only_with_analyzer	1
parallel_replicas_insert_select_local_pipeline	1
parallel_replicas_connect_timeout_ms	300
session_timezone	``
low_priority_query_wait_time_ms	1000
max_limit_for_vector_search_queries	1000
hnsw_candidate_list_size_for_search	256
vector_search_filter_strategy	auto
vector_search_postfilter_multiplier	1
allow_experimental_delta_kernel_rs	1

Experimental settings

Name	Default
allow_experimental_kafka_offsets_storage_in_keeper	0
allow_experimental_correlated_subqueries	0
allow_experimental_materialized_postgresql_table	0
allow_experimental_funnel_functions	0
allow_experimental_nlp_functions	0
allow_experimental_hash_functions	0
allow_experimental_object_type	0
allow_experimental_time_series_table	0
allow_experimental_vector_similarity_index	0
allow_experimental_codecs	0
throw_on_unsupported_query_inside_transaction	1
wait_changes_become_visible_after_commit_mode	wait_unknown
implicit_transaction	0
grace_hash_join_initial_buckets	1
grace_hash_join_max_buckets	1024
join_to_sort_minimum_perkey_rows	40
join_to_sort_maximum_table_rows	10000
allow_experimental_join_right_table_sorting	0
allow_statistics_optimize	0
allow_experimental_statistics	0
allow_experimental_inverted_index	0
allow_experimental_full_text_index	0
allow_experimental_lightweight_update	0
allow_experimental_join_condition	0
allow_experimental_live_view	0
live_view_heartbeat_interval	15
max_live_view_insert_blocks_before_refresh	64
allow_experimental_window_view	0
window_view_clean_interval	60
window_view_heartbeat_interval	15
wait_for_window_view_fire_signal_timeout	10
stop_refreshable_materialized_views_on_startup	0
allow_experimental_database_materialized_postgresql	0
allow_experimental_query_deduplication	0
allow_experimental_database_iceberg	0
allow_experimental_database_unity_catalog	0
allow_experimental_database_glue_catalog	0
allow_experimental_database_hms_catalog	0
allow_experimental_kusto_dialect	0
allow_experimental_prql_dialect	0
enable_adaptive_memory_spill_scheduler	0
make_distributed_plan	0
distributed_plan_execute_locally	0
distributed_plan_default_shuffle_join_bucket_count	8
distributed_plan_default_reader_bucket_count	8
distributed_plan_force_exchange_kind	``
allow_experimental_ts_to_grid_aggregate_function	0
allow_experimental_replacing_merge_with_cleanup	0
allow_experimental_reverse_key	0
enable_replacing_merge_with_cleanup_for_min_age_to_force_merge	0
force_read_through_cache_for_merges	0
merge_selector_algorithm	Simple
notify_newest_block_number	0
part_moves_between_shards_delay_seconds	30
part_moves_between_shards_enable	0
remote_fs_zero_copy_path_compatible_mode	0
remote_fs_zero_copy_zookeeper_path	/clickhouse/zero_copy
remove_rolled_back_parts_immediately	1
shared_merge_tree_enable_coordinated_merges	0
shared_merge_tree_enable_keeper_parts_extra_data	0
shared_merge_tree_merge_coordinator_election_check_period_ms	30000
shared_merge_tree_merge_coordinator_factor	2
shared_merge_tree_merge_coordinator_fetch_fresh_metadata_period_ms	10000
shared_merge_tree_merge_coordinator_max_merge_request_size	20
shared_merge_tree_merge_coordinator_max_period_ms	10000
shared_merge_tree_merge_coordinator_merges_prepare_count	100
shared_merge_tree_merge_coordinator_min_period_ms	1
shared_merge_tree_merge_worker_fast_timeout_ms	100
shared_merge_tree_merge_worker_regular_timeout_ms	10000