Sqoop Import Fails

I'm trying to import a Table from Oracle to Hive using Sqoop. I used the following command:

sqoop-import --connect jdbc:<connection> --table test1 --username test --password test --hive-table hive_test --create-hive-table --hive-import -m 1

But this gives me the error

 Encountered IOException running import job: org.apache.hadoop.mapred.FileAlreadyExistsException: Output directory <hdfs path> already exists

So I read online in many forums and it said that I should delete the directory and run the command again. I did exactly that, but I still keep getting the error.

Solution

You need to understand working of Sqoop hive Import.

Import data to HDFS <some-dir>
Create hive table <some-table> IF NOT EXISTS
LOAD data inpath '<some-dir>' into table <some-table>

You are getting the error at step 1.

Output directory <hdfs path> already exists

Delete this <hdfs path> and proceed.

Better way:

No need to delete this manually everytime.

Use --delete-target-dir in the command. It will

Delete the import target directory if it exists

P.S. No need to use --create-hive-table with --hive-import. --hive-import by default create table for you.