doris2.1.4 非分区表自动分桶的规则

Viewed 69

doris 2.1.4 版本
建表语句 CREATE TABLE test (
id BIGINT NULL COMMENT 'id',
name INT NULL COMMENT '名称',

) ENGINE=OLAP
UNIQUE KEY(id)
COMMENT '测试表'
DISTRIBUTED BY HASH(id) BUCKETS AUTO
PROPERTIES (
"replication_allocation" = "tag.location.default: 3",
"min_load_replica_num" = "-1",
"is_being_synced" = "false",
"storage_medium" = "hdd",
"storage_format" = "V2",
"inverted_index_storage_format" = "V1",
"estimate_partition_size" = "1G",
"enable_unique_key_merge_on_write" = "true",
"light_schema_change" = "true",
"disable_auto_compaction" = "false",
"enable_single_replica_compaction" = "false",
"group_commit_interval_ms" = "10000",
"group_commit_data_bytes" = "134217728"
);

show data from 表; 显示34G数据

show partitions from 表;显示 2个buckets

辛苦帮忙看下正常吗?或者我们应该怎么使用自动分桶的功能呢?谢谢

1 Answers

自动分桶是按照estimate_partition_size 和历史分区的数据来算出 bucket 的数量,开始的时候建表没历史数据 ,所以不能保证每个 partition 都是这么大。