Five Stars

Column data mismatch in hive

Dear folks,

Present I am working on talend big data 6.4.1v.

I am ingesting data from oracle to hive by using tHdfsOutput, tHiveLoad components.

By using tHiveCreateTable component I was created table in Hive by using field delimiter comma(,).

After ingestion completed, verified records in hive there is some values jump from one column to another.

I replaced field seperator with "\t" , "I" after that also I am facing some mismatch column values.

Kindly suggest me which field seperator is better for me to avoid this issue.

 

 

Regards,

Rupesh.M

1 REPLY
Five Stars

Re: Column data mismatch in hive

Hi Rupesh,

It sounds like your source data contains a lot of open text fields (such as comments on a record) - often these may contain pipes due to end users catching them while entering data.

 

It's a better option to sanitise the input rather than try to work around it. Using the tReplace component on the columns that can receive open text means you can replace special characters with a space or blanks. You could use a regex specifying which characters are acceptable (for example [a-zA-Z0-9]+ for only alphanumeric characters) so the rest get replaced with a space