问题描述
我使用 Clickhouse 数据库。
我有分段输入的文本。我需要将所有文本分组在一行中，而没有重复的段
当count(segment) = total Segment。
我创建了第一个表来插入段。 然后我创建一个 MATERIALIZED 视图来删除重复的行并对所有段进行分组。
然后将其移动到第二个表，但我的问题是视图不显示值。 当我运行查询时它工作正常
第一个表：
CREATE table test_table_1
(
id UInt16,text String,totalSegment UInt16,segmentNumber UInt16,reference UInt16,insertTime DateTime
)
ENGINE = ReplacingMergeTree()
order by (id,text,totalSegment,segmentNumber,reference );
第二个表：
CREATE table test_table_2
(
message String,totalSegment UInt16
)
ENGINE = MergeTree()
order by (reference );
实体化视图：
CREATE MATERIALIZED VIEW consumer To test_table_2
as
SELECT groupArray(text) as message,reference,totalSegment
from (
select *
from
(
SELECT COUNT(*) as countSegment,reference
from test_table_1
where (text,insertTime) in (
select text,max(insertTime) as insertTime
from test_table_1
group by text,reference
)
group by totalSegment,reference
HAVING count(*) = totalSegment
)as a
left join (
select text,max(insertTime) as insertTime
from test_table_1
group by text,reference
)as b using (reference)
order by segmentNumber
)
grousp by reference,totalSegment
order by reference ;
插入数据：
INSERT INTO test_table_1 (id,insertTime) VALUES(1,'A1 ',3,1,'101',Now() );
INSERT INTO test_table_1 (id,insertTime) VALUES(2,'B1 ',2,'202',insertTime) VALUES(3,'C2 ','303',insertTime) VALUES(4,'A3 ',insertTime) VALUES(5,'A2 ',insertTime) VALUES(6,'C1 ',insertTime) VALUES(7,'C3 ',insertTime) VALUES(8,'B2 ',Now() );
解决方法
MV 是插入触发器。它从不读取源表 -- test_table_1 !!!
https://den-crane.github.io/Everything_you_should_know_about_materialized_views_commented.pdf