Python is a popular programming language that can be used for advanced analytics and data modeling. Dundas BI integrates with Python at the data cube level via two transforms, which you can use to supply data to the data cube or perform data processing/analysis on other data.
2. Install Python
If necessary, download and run the installer for the 64-bit version of Python on your Dundas BI server with the correct version number:
- For Dundas BI version 7 and above installed on a Windows server, Python is installed as a prerequisite when creating an instance. Check the system requirements for your version of Dundas BI for more details.
- For earlier versions of Dundas BI installations, or for Dundas BI installed on a Linux server or via Docker images, ensure Python 3.6 is installed for all users (--enable-shared on Linux) to use in Python transforms. On Linux, the terminal command python3 --version should display version 3.6.
The installation process on a Windows server will be similar to the following:
Select Customize installation and click Next on the Optional Features page.
Select Install for all users on the Advanced Options page and click Install.
3. Install packages
Packages must also be installed by an administrator on the server that runs Dundas BI before they can be used in Python transforms.
For example, to install the common pandas package, which includes other packages such as NumPy:
- Open the command prompt or terminal on the Dundas BI server as an administrator (e.g., right-click on Command Prompt in the start menu and choose Run as administrator).
- You may need to navigate to the particular Scripts folder corresponding with the version of Python used by Dundas BI. For example:
cd C:\Program Files\Python38\Scripts
- Use pip to install the package. For example:
pip install pandas
- The Python-related transforms in Dundas BI require a result to be returned by the Python script, which can be represented as a table. This includes results generated using the NumPy and Pandas packages for data analysis.
- You may be required to restart the server after installing Python.