Raphael Nguyen created SQOOP-3207:
-------------------------------------
Summary: Codegen is using column labels rather than column names
Key: SQOOP-3207
URL: https://issues.apache.org/jira/browse/SQOOP-3207
Project: Sqoop
Issue Type: Bug
Components: codegen
Affects Versions: 1.4.5
Environment: * Red Hat Enterprise Linux Server release 6.2 (Santiago)
* Teradata 14.10.07.21
* Sqoop 1.4.6
Reporter: Raphael Nguyen
Priority: Minor
We are using codegen to create external Hive tables.
Sqoop 1.4.4 was using getColumnName() in *SqlManager.java* and
*ResultSetPrinter.java*. It looks like
[SQOOP-585|https://issues.apache.org/jira/browse/SQOOP-585] does the opposite,
using getColumnLabel() first.
The issue can be seen when column labels (aliases) contain spaces, which is
often the case for Teradata sources. When running codegen, the resulting Hive
create statement will persist the spaces in the column labels as column names.
The desired behavior is to use the column names instead.
*Example:*
{code:java}
CREATE TABLE IF NOT EXISTS `null` ( `FIRST NAME` STRING, `LAST NAME` STRING)
COMMENT 'Imported by sqoop on 2017/06/28 14:15:35' ROW FORMAT DELIMITED FIELDS
TERMINATED BY '\001' LINES TERMINATED BY '\012' STORED AS TEXTFILE;
{code}
*Sqoop 1.4.4 Example:*
{code:java}
CREATE TABLE IF NOT EXISTS `null` ( `FIRST_NAME` STRING, `LAST_NAME` STRING)
COMMENT 'Imported by sqoop on 2017/06/28 14:15:35' ROW FORMAT DELIMITED FIELDS
TERMINATED BY '\001' LINES TERMINATED BY '\012' STORED AS TEXTFILE;
{code}
*Codegen example:*
{code:java}
sqoop codegen -libjars ../lib/tdgssconfig-1.5.1.jar --driver
com.teradata.jdbc.TeraDriver --connect
jdbc:teradata://my_hostname/DATABASE=my_db --username my_username --password
my_password --outdir /users/raphnguyen/my_table/src_generated --bindir
/users/raphnguyen/my_table/jar_generated --query "SELECT * FROM MY_DB.MY_TABLE
WHERE \$CONDITIONS" --hive-import
{code}
Related: [SQOOP-585|https://issues.apache.org/jira/browse/SQOOP-585]
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
