Details
Major Description
Carbondata1.3.1 with Spark2.1.2,after insert into values,SELECT is right.but next day,data of SELECT is null but the line number is right.
create table carbon01(id int,name int,age int,sex int) stored by 'carbondata';
insert into carbon01 values(1,1,1,1);
select * from carbon01;
1 1 1 1
then,i exit spark-sql.
but,next day,
select * from carbon01;
null null null null
.it may not because of time.
after one line turn to null,i insert into again.
insert into carbon01 values(2,2,2,2);
then i select,the result is:
null null null null
2 2 2 2
the log is:
ResultCollectorFactory:[Executor task launch worker for task 1][partition...] Restructure dictionary vector collector is used to scan and collect the data
ResultCollectorFactory:[Executor task launch worker for task 2][partition...] Vector based dictionary collector is used to scan and collect the data
and i follow source code,in task1, tableBlockInfo.getDetailInfo.columnSchemas.get(0).isDimensionColumn=true;
in task2, tableBlockInfo.getDetailInfo.columnSchemas.get(0).isDimensionColumn=false;