[SQOOP-3290] Not able to import from SQL Server on HDFS Parquet (Date columns)

• Type: Test
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Assignee: Unassigned
• Reporter: Ashish Kumar Sinha
• Created: 23/Feb/18 13:47
• Updated: 23/Feb/18 13:47

Description

Hi,

I am able to import data from SQL Server to HDFS Parquet format, but the script fails when there is date or datetime or datetime2 column in SQL Server table. 

I tried to use --map-column-hive option as mentioned below

sqoop import --connect "jdbc:jtds:sqlserver://<<Servername>>;useNTLMv2=true;domain=APNET;databaseName=<<DB Name>>" \
-connection-manager org.apache.sqoop.manager.SQLServerManager --driver net.sourceforge.jtds.jdbc.Driver --username *** --password *** --table T1 --split-by C1-target-dir /user/user1/bkp1/11/ --map-column-java D1=timestamp --as-parquetfile – --schema=sch1;

 

The script succeeds successfully but when I load the file in Hive parquet table, the date column is coming as NULL. Am I doing in wrongly or is it a known bug for sqoop for SQL conversion error? Please advice.