Exporting data from Hadoop

Depending on what type of data you plan to transfer from Hadoop, and whether the data on Hadoop was imported from Db2® Warehouse, you can follow one of the three export scenarios.

About this task

The scenarios are described in detail in the following sections.
  1. You can export data that was previously imported from Db2 Warehouse, as described in Exporting previously imported files
  2. You can export text files that were not previously imported from Db2 Warehouse, as described in Exporting text files from Hadoop.
  3. You can export a Hive table (in any format) that was not imported from Db2 Warehouse, as described in Exporting a Hive table that was not imported from Db2 Warehouse.
All of the mentioned export types can be run from Hadoop - using the fdm.sh script.
Note: When exporting new text data from Hive, you can either choose to export text files (scenario 2), or to export a table (scenario 3). Exporting a text file is better from performance point of view, however it requires more information about the data (table structure, separator, null representation etc.), and an additional manual step of creating the table in Db2 Warehouse.