[SQOOP-2664] Duplicate records found when split-by column is of type char(n)

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.4.5
• Component/s: sqoop2-jdbc-connector
• Labels: import split-by sqoop

Environment

Hortonworks: 2.2.4.2
Sqoop: 1.4.5
MS SQL: R2008

• External issue URL: https://issues.apache.org/jira/browse/SQOOP-2536
• Assignee: Unassigned
• Reporter: Dhaval Modi
• Created: 05/Nov/15 23:08
• Updated: 26/Nov/17 16:11

Description

Hi,

While working with SQOOP, we found an issue where records are duplicated while importing the data. This is seen when split-by column is of type char.

We understand, ideally, integral columns with not null type must be chosen for split-by column, but in our case, all integral columns has null values. There is open bug (SQOOP - 2536) raised for this issue.

But as sqoop have support for char datatype to be used in split-by column, and it giving unexpected results. Thus raising this bug.

Issue Links

• is related to: SQOOP-3263 Duplicate rows found when split-by column is of textual type due to different charset difference of sqoop and hadoop