Support Questions

Find answers, ask questions, and share your expertise

Ambari Hive View output skips every 101st row

avatar

**** Update: This has been fixed in Ambari 2.5 (released April 4, 2017). See below for more details. ****

This has been observed in Ambari 2.4 using HDP 2.5.0/HDP 2.5.3.4-5.

Customer is having the issue and I am able to replicate it.

Using the Ambari Hive view, I'm running a simple select query. By using tpcds data, it is easy to spot the problem since the customer table has an auto-number row_id.

select c_customer_sk, c_customer_id, c_last_name from customer limit 204;

The query runs successfully & the results page in the browser every 50 rows.

  • The first page shows rows 1..50
  • the second page shows rows 51..100,
  • but the 3rd page starts with row 102. Row 101 is skipped on the output.

13055-screen-shot-2017-02-27-at-53740-pm.png

  • The 3rd page shows rows 102..151
  • The 4th page shows rows 152..201
  • but the 5th page starts with row 203. Row 202 is skipped on the output.

13039-screen-shot-2017-02-27-at-53100-pm.png

The same pattern is displayed when using "Save Results" -> "Download as CSV" - every 101st row is skipped.

13056-screen-shot-2017-02-27-at-54145-pm.png

In Beeline when I run the same query - no issues. All rows are displayed.

13054-screen-shot-2017-02-27-at-53535-pm.png

Have attempted to set the following values in Ambari but they don't affect the results:

  • set hive.cli.print.header=true;
  • set showHeader false (beeline)
  • set headerInterval 0

Thanks in advance for the assistance.

1 ACCEPTED SOLUTION

avatar

@bpreachuk

This is a bug and it was fixed in 2.4.x version but still not part of apache. If you are hortonworks customer you can request support for HOTFIX.

View solution in original post

5 REPLIES 5

avatar

@bpreachuk

This is a bug and it was fixed in 2.4.x version but still not part of apache. If you are hortonworks customer you can request support for HOTFIX.

avatar

Thanks @apappu. I will request the hotfix for the customer.

avatar
New Contributor

@bpreachuk @apappu Hey - I am having the same issue on Ambari using HDP 2.5 when looking to export results to csv format. Is there a workaround I could use to fix this?

Thanks!

avatar

Hi @Ciarán Porter. There is a hotfix required for this issue. In the meantime you can use the beeline cli and send the output to a csv format. A good explanation is found here: https://community.hortonworks.com/questions/25789/how-to-dump-the-output-from-beeline.html

avatar

Update: This has been fixed in Ambari 2.5 (released April 4th, 2017). An added bonus is that Ambari 2.5 includes the new Hive View 2.0 - a vastly improved GUI for developing Hive apps. See attached picture:

14389-screen-shot-2017-04-05-at-55852-am.png