[
https://issues.apache.org/jira/browse/HUDI-5857?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jing Zhang updated HUDI-5857:
-----------------------------
Description:
Snapshot query result is wrong after apply insert overwrite to an existed table
with simple bucket index.
The bug could be produced by the following steps.
# create a mor table with bucket index
{code:java}
create table test_hudi_zj0221(
id int,
name string,
price double,
ts long,
dt string) using hudipartitioned by (dt)
options(
type='mor',
primaryKey='id',
preCombineField = 'ts',
'hoodie.index.type'='BUCKET',
'hoodie.bucket.index.num.buckets'='8) {code}
# insert into data
{code:java}
insert
into test_hudi_zj0221 select 8 as id, 'hudi3' as name, 30 as price, 3000 as ts,
'2021-05-05' as dt;
insert
into test_hudi_zj0221 select 9 as id, 'hudi3' as name, 30 as price, 3000 as ts,
'2021-05-05' as dt;
insert
into test_hudi_zj0221 select 10 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 11 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 12 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 13 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 14 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 15 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt; {code}
# find something wrong, use insert overwrite to overwrite a partition
{code:java}
insert overwrite table test_hudi_zj0221 partition(dt = '2021-05-05') select
2222, 'a2',30, 3000; {code}
# snapshot query on the table
{code:java}
select * from test_hudi_zj0221 where dt='2021-05-05';
-- or
select * from test_hudi_zj0221; {code}
was:
Snapshot query result is wrong after apply insert overwrite to an existed table
with simple bucket index.
The bug could be produced by the following steps.
# create a mor table with bucket index
{code:java}
create table test_hudi_zj0221(
id int,
name string,
price double,
ts long,
dt string) using hudipartitioned by (dt)
options(
type='mor',
primaryKey='id',
preCombineField = 'ts',
'hoodie.index.type'='BUCKET',
'hoodie.storage.layout.partitioner.class'='org.apache.hudi.table.action.commit.SparkBucketIndexPartitioner',
'hoodie.bucket.index.num.buckets'='8',
'hoodie.datasource.write.recordkey.field' = 'id',
'hoodie.storage.layout.type'='BUCKET') {code}
# insert into data
{code:java}
insert
into test_hudi_zj0221 select 8 as id, 'hudi3' as name, 30 as price, 3000 as ts,
'2021-05-05' as dt;
insert
into test_hudi_zj0221 select 9 as id, 'hudi3' as name, 30 as price, 3000 as ts,
'2021-05-05' as dt;
insert
into test_hudi_zj0221 select 10 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 11 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 12 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 13 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 14 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt;
insert
into test_hudi_zj0221 select 15 as id, 'hudi3' as name, 30 as price, 3000 as
ts, '2021-05-05' as dt; {code}
# find something wrong, use insert overwrite to overwrite a partition
{code:java}
insert overwrite table test_hudi_zj0221 partition(dt = '2021-05-05') select
2222, 'a2',30, 3000; {code}
# snapshot query on the table
{code:java}
select * from test_hudi_zj0221 where dt='2021-05-05';
-- or
select * from test_hudi_zj0221; {code}
> Snapshot query result is wrong after apply insert overwrite to an existed
> table with simple bucket index
> --------------------------------------------------------------------------------------------------------
>
> Key: HUDI-5857
> URL: https://issues.apache.org/jira/browse/HUDI-5857
> Project: Apache Hudi
> Issue Type: Bug
> Reporter: Jing Zhang
> Assignee: Jing Zhang
> Priority: Major
> Labels: pull-request-available
> Attachments: image-2023-02-27-15-55-26-568.png
>
>
> Snapshot query result is wrong after apply insert overwrite to an existed
> table with simple bucket index.
> The bug could be produced by the following steps.
> # create a mor table with bucket index
> {code:java}
> create table test_hudi_zj0221(
> id int,
> name string,
> price double,
> ts long,
> dt string) using hudipartitioned by (dt)
> options(
> type='mor',
> primaryKey='id',
> preCombineField = 'ts',
> 'hoodie.index.type'='BUCKET',
> 'hoodie.bucket.index.num.buckets'='8) {code}
> # insert into data
> {code:java}
> insert
> into test_hudi_zj0221 select 8 as id, 'hudi3' as name, 30 as price, 3000 as
> ts,
> '2021-05-05' as dt;
> insert
> into test_hudi_zj0221 select 9 as id, 'hudi3' as name, 30 as price, 3000 as
> ts,
> '2021-05-05' as dt;
> insert
> into test_hudi_zj0221 select 10 as id, 'hudi3' as name, 30 as price, 3000 as
> ts, '2021-05-05' as dt;
> insert
> into test_hudi_zj0221 select 11 as id, 'hudi3' as name, 30 as price, 3000 as
> ts, '2021-05-05' as dt;
> insert
> into test_hudi_zj0221 select 12 as id, 'hudi3' as name, 30 as price, 3000 as
> ts, '2021-05-05' as dt;
> insert
> into test_hudi_zj0221 select 13 as id, 'hudi3' as name, 30 as price, 3000 as
> ts, '2021-05-05' as dt;
> insert
> into test_hudi_zj0221 select 14 as id, 'hudi3' as name, 30 as price, 3000 as
> ts, '2021-05-05' as dt;
> insert
> into test_hudi_zj0221 select 15 as id, 'hudi3' as name, 30 as price, 3000 as
> ts, '2021-05-05' as dt; {code}
> # find something wrong, use insert overwrite to overwrite a partition
> {code:java}
> insert overwrite table test_hudi_zj0221 partition(dt = '2021-05-05') select
> 2222, 'a2',30, 3000; {code}
> # snapshot query on the table
> {code:java}
> select * from test_hudi_zj0221 where dt='2021-05-05';
> -- or
> select * from test_hudi_zj0221; {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
