Uninstalling and reinstalling wheel on cluster running DBR 8.3 may fail
If you have a named cluster specification in your Databricks environment and it had the current or a previous build of the datagenerator installed, when you uninstall the library and reinstall it , it may fail
Expected Behavior
Uninstall followed by reinstall should succeed
Current Behavior
Uninstall followed by re-install may fail.
Workaround
- make sure the wheel does not have a name like
dbldatagen-0.2.0rc1-py3-none-any.whl (1)which may result from multiple downloads on the same machine - dont use a saved cluster definition - use a new cluster definition
Our plan is to move to a PIP based install which should make installation easier
Your Environment
dbldatagenversion used: release candidate 2- Databricks Runtime version: Databricks 8.3
- Cloud environment used: Azure
You can also install in a Databricks notebook environment using the %pip command in a notebook cell. The %pip command does not illustrate the issue above.
To install the library via %pip, add a cell with the following content at the start of the notebook and execute it:
%pip install git+https://github.com/databrickslabs/dbldatagen
Newest versions will be based on Databricks runtime 9.1 LTS or later