建表** ****** ***** study_livelog (uid UInt**,platform Stng,app_id Int16,channel Int16,category_id Int32,subject_id Int32,media_type Int8,media_id UInt,media__id Int,d Stng,sp UInt32,ep UInt32,st DateTime,et DateTime,sr float,tt Int32,ip Stng,version Stng,device Stng,kernel Stng,address Stng,updated_date DateTime,watch_duration UInt32) ENGINE = ReplicatedReplacin*ergeTree(’/clickhouse/study/study_livelog’, ‘{replica}’, watch_duration) partition by toYYYYMM(st) order by (uid, media_type, media_id,st); 在表的数据目录下，会产生大量的新目录，每入一条数据会产生一个目录，如下图。请问这是什么原因？