topic Re: Enable programmatically writing to files in Machine Learning

Enable programmatically writing to files

rubenteixeira — Mon, 23 Jan 2023 16:32:53 GMT

Hi everyone. I'm training a time series forecasting model in Azure Databricks. When I try to parallelize, it gives me this error:

I have Contributor permission on the Databricks service, on the Azure Portal, and I'm an admin inside the Databricks workspace. What permissions am I missing?

Re: Enable programmatically writing to files

Hubert-Dudek — Mon, 23 Jan 2023 17:31:48 GMT

Can't you use data frames API? Using RDD directly is an outdated method as it is not optimized. Dataframe is automatically distributed (parallelized using partitions and cores). Additionally, how big is a dataset? as running collect() is also a killer and should not be used in production or to return more than a dozen rows.

Re: Enable programmatically writing to files

jose_gonzalez — Fri, 24 Feb 2023 23:38:52 GMT

Hi @Rúben Teixeira,

Just a friendly follow-up. Did any of the responses help you to resolve your question? if it did, please mark it as best. Otherwise, please let us know if you still need help.

Re: Enable programmatically writing to files

Debayan — Tue, 24 Jan 2023 07:23:50 GMT

Hi,

The error says the below filesystem is read-only and not RW.