Hello Markus, Can I ask you one question? Does that make any change? Both are for the same purpose, right? Please correct me if I am wrong.
On Sun, Dec 3, 2017 at 10:39 AM, Markus Kemper <[email protected]> wrote: > Hello Ajana, > > Have you tried using (--input-lines-terminated-by '\n') instead of > (--lines-terminated-by > '\n')? > > > Markus Kemper > Customer Operations Engineer > [image: www.cloudera.com] <http://www.cloudera.com> > > > On Sun, Dec 3, 2017 at 8:45 AM, Ajana Chandiruthil Sathian < > [email protected]> wrote: > >> >> ---------- Forwarded message ---------- >> From: Ajana Chandiruthil Sathian <[email protected]> >> Date: Wed, Nov 15, 2017 at 11:09 AM >> Subject: Upsert option in Sqoop export command. >> To: [email protected] >> >> >> To whom ever it may concern, >> >> I have a csv file in Hadoop and I did Sqoop export to Oracle. The column >> data type order in ODS is number,date, date,float,varchar and the column >> data type in the csv file is number,date, float,date,varchar and I used the >> --columns sqoop argument to get control in column ordering and it worked. >> But I could not control the column ordering when I was doing the upsert >> operation in sqoop export. It is giving me misalignment in ODS after >> Sqoop( please see the attached image). The below given is the sqoop command: >> >> sqoop export --connect ConnectionString \ >> --username xxx \ >> --password xxxx \ >> --table tableName \ >> --export-dir /dir/TestUpdate.txt \ >> --input-fields-terminated-by ',' \ >> --lines-terminated-by '\n' \ >> --update-key column_name \ >> --update-mode allowinsert \ >> --columns "id,START_DT,VAL,end_DT,QUALITY" \ >> -m 4 >> >> >
