# Cognite DB Extractor

The Cognite DB Extractor is a general database extractor that connects to one or more databases, executes queries, and sends the results to the staging area in Cognite Data Fusion (CDF).

The extractor connects to databases using ODBC, and you need ODBC drivers for the databases you're connecting to. Typically, the database vendors provide the necessary ODBC drivers for their databases.

The extractor is available as:

  • A native Windows application.
  • A Docker image for Unix-based systems (for example, Mac OS and Linux).

# Server requirements

You can install the extractor on a physical server or on a virtual machine on-premises or in the cloud.

Hardware requirements Software requirements (for Windows version)
The minimum requirements are:
  • At least a 2 core CPU
  • 1 GB RAM*
  • 1 GB HD for logging and storage
  • Windows 10
  • ODBC driver(s) for your database(s)**

* You can adjust the configuration the reduce the memory usage. The minimum requirement if for the recommended default values.

** The Docker image includes drivers for Postgres, Microsoft SQLServer, MySQL, and Oracle DB.

# Prerequisites

To configure and deploy the Cognite DB Extractor, you also need:

  • Network access to the database and the CDF project (firewall openings).
  • User credential providing the necessary access to the database.
  • An API-key for the CDF project with the following capabilities:
    • raw:read
    • raw:write
    • raw:list

# Incremental load

The Cognite DB Extractor supports incremental load (processing only changes since the previous run) if it can use a table column to determine when a row was added or changed. Examples are a timestamp for the last edit of the row, a row index (assuming immutable rows), or similar.

Last Updated: 6/20/2020, 6:57:42 AM