I want to create a sklearn model on azure machine learning service. But when I run the script then I got error in pyodbc. How to install pyodbc on azure machine learning service?
Asked
Active
Viewed 1,379 times
0
-
Could you please add some more information to this question? It would be good to know how you have set up your environment and what type of error message you get. – Arvid Bäärnhielm Jan 02 '19 at 07:13
-
Did you read this? https://stackoverflow.com/questions/48885715/how-do-you-use-pyodbc-in-azure-machine-learning-workbench – user1965074 Feb 16 '19 at 17:18
-
If you are connecting to a MS SQL Server database, you can use `azureml.dataprep`, as outlined in the [documentation](https://learn.microsoft.com/en-us/azure/machine-learning/service/how-to-load-data#load-sql-data) or in [this SO answer](https://stackoverflow.com/a/54692648/4240413), without the need pyodbc. – Davide Fiocco Apr 07 '19 at 13:36
1 Answers
2
Maybe you already solved your issue, but I ran into the same problem with the docker image from Microsoft in Azure Machine Learning Services. (mcr.microsoft.com/azureml/base)
So I built my own docker image based on the base image, and installed the missing MSSQL driver for PyOdbc and some additional locales required by the driver.
hypnobrew/azure-ml:0.1
Hypnobrew
- 1,120
- 9
- 23
-
I have used `sql_driver = "SQL Server"` Sql driver. Can you please gitve me the breaf explanation about how to solve the above issue. Can you please send me the Pyodbc connection code snippent? – Mar 20 '19 at 09:15
-
-
Yes, by running your experiments in a docker container with the pyodbc driver installed. Like my docker image above. – Hypnobrew Mar 21 '19 at 16:19
-
Otherwise I do not think Microsoft support custom libraries to depend on, in their Azure maching learning services. (Added github link to the Docker Hub image page) – Hypnobrew Mar 22 '19 at 07:22