[jira] [Created] (FLINK-10673) Table API / SQL UIDs not the only one

["Fan weiwen (JIRA)" <ji...@apache.org>]

Fan weiwen created FLINK-10673:
----------------------------------

             Summary: Table API / SQL UIDs not the only one
                 Key: FLINK-10673
                 URL: https://issues.apache.org/jira/browse/FLINK-10673
             Project: Flink
          Issue Type: Bug
          Components: Table API & SQL
    Affects Versions: 1.6.1, 1.5.4
         Environment: flink 1.5.0
            Reporter: Fan weiwen


a job  have two sql 
source is kafka 
sink is redis  or other sink 

Asql 


{code:java}
//代码占位符
select 
  reqIp as factorContenta, 
  count(*) as eCount, 
  60 * 60 as expire 
from 
  kafka_source 
where 
  uri is not null 
group by 
  hop( 
    rowtime, 
    interval '2' second, 
    interval '60' minute 
  ), 
  reqIp 
{code}



Bsql 


{code:java}
//代码占位符
select 
        uid as factorContentb, 
  count(*) as eCount, 
  60 * 60 as expire 
from 
  kafka_source 
where 
  uri is not null 
group by 
  hop( 
    rowtime, 
    interval '2' second, 
    interval '60' minute 
  ), 
  uid 
{code}



now only start Asql  stop Bsql    sink  have key   656.19.173.34 

then stop Asql and savepoint hdfs   now  del key 656.19.173.34( if sink is 
kafka Don't delete) 

start Bsql  from savepoint 
you will find sink have key   656.19.173.34 and 6630519 all exist 

Bsql fetch Asql savepoint result 

i think sql uids not the only one 

Who can help me see this problem? 

my test data is 


{code:java}
//代码占位符
{ 
   "reqIp" : "656.19.173.34", 
   "rowtime" : 1537950912546, 
   "uid" : 6630519, 
   "uri" : "/web" 
} 
{code}



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)