1 Reply Latest reply: Jul 13, 2016 11:25 AM by Manasa kamma RSS

    Connecting to hive and loading the script works for some tables, but for one particular table i see an error

    Manasa kamma

      Hi,

      I'm new to qlikview and 'm trying to connect to hive using microsoft odbc hive drivers from qlik.

      Connecting to hive and loading the script works for some tables,but for one particular table i see below error

      'QVX_UNEXPECTED_END_OF_DATA: SQL##f - SqlState: S1000, ErrorCode: 35, ErrorMsg: [Microsoft][HiveODBC] (35) Error from Hive: error code: '0' error message: 'java.io.IOException: java.lang.RuntimeException: serious problem'.'

      This used to work fine some days back,but since couple of days it is failing continuously.

      There are no duplication with column names, even when i try to load the script for just 1 column and just for first 10 rows i still get the same error. when i include a where clause saying 1=2 the it fetches 0 records without any error.

      i spent hours debugging this .Kindly let me know if anyone of you came across similar error and solution that worked.

       

      Many thanks,

      ManasaK

        • Re: Connecting to hive and loading the script works for some tables, but for one particular table i see an error
          Manasa kamma

          Hi,

          I tried all the possible scenarios to debug it and got some clue about the issue mentioned above.

          1)Did a Select * from table1 where 1=2; (fetches 0 records without error)

                       First 1

                      Select column1 from table1 ;(Tried each column individually to fetch single row ,it fails showing ODBC read failed)

          Here figured out issue is not with table structure and it fails while it tries to read any data from table.

          2)To check it was issue with ODBC connector tried using OLEDB to hive and still same issue.

          Here figured out issue is not with connector.

          3)Downloaded data into CSV and tried to load and it worked perfectly.

          Here figured out issue is not with type of data in table.

          4)Created a similar table using

          Create table table1_test(select * from table1); it worked perfectly when i load data from table_test.

          Here figured out issue is in not with table structure and data in it but with Constraints or any other approach followed while creating table.

          5)compared create table scrtipt for table1 and table1_test by using.

          show create table table1_test;

          Show create table table1;

          The only difference i see was table1 has partition and table1_test doesn't have partition.

          6)So i tried to load table2 which also has partition  and  it worked perfectly.

          So compared table1 and table2 create script and difference i see

          table2 has partion along with cluster by(Column) and table1 has partition alone.

          Now i'm trying to figure out how is cluster by responsible for the read failed error.

          Please do share your thoughts if you know how partition with cluster by and partition with out cluster by is making this difference?

           

          Hope this helps some of you facing similar issues.

           

          Thanks,

          Manasak