Adding a dataset from HDFS

You can access data stored on HDFS (Hadoop File System), directly from the Talend Data Preparation interface and import it in the form of a dataset.

Procedure

In the Datasets view of the Talend Data Preparation homepage, click the white arrow next to the Add Dataset button.
Select HDFS.

The Add an HDFS dataset form opens.
In the Dataset name field, enter the name you want to give your dataset.
In the User name field enter your Linux user name.

This user must have the reading rights on the file that you want to import.
To enable Kerberos authentication, select the Use Kerberos check box.
In the Principal
In the Keytab file field, enter the location of your keytab file.
The keytab file must be accessible by the Spark Job Server.
You can manually configure Talend Data Preparation to display a default value in those fields.
In the Format field, select the format that corresponds to the file that you want to import.
For HDFS files, Talend Data Preparation supports CSV, AVRO and PARQUET.
Information noteWarning: Talend Data Preparation does not support the import of PARQUET files that contain data with the INT96 type. We recommend adjusting your source file if that is the case.

If you choose CSV, select the record and field delimiter, as well as the text enclosure and escape character, and the encoding for the file you want to import.
In the Path field, enter the complete URL of your file in the Hadoop cluster.
Click the Add Dataset button.

The data extracted from the cluster directly opens in the grid and you can start working on your preparation.

The data is still stored in the cluster and doesn't leave it, Talend Data Preparation only retrieves a sample on-demand.

Your dataset is now available in the Datasets view of the application home page.

If you find any issues with this page or its content – a typo, a missing step, or a technical error – please let us know!