I have a simple regression task (using a LightGBMRegressor) where I want to penalize negative predictions more than positive ones. Is there a way to achieve this with the default regression LightGBM objectives (see https://lightgbm.readthedocs.io/en/latest/Parameters.html)? If not, is it somehow possible to define (many example for default LightGBM model) and pass a custom regression objective?

This is to track implementation of the ML-Features: https://spark.apache.org/docs/latest/ml-features

Bucketizer has been implemented in dotnet/spark#378 but there are more features that should be implemented.

• Feature Extractors
  • TF-IDF
  • Word2Vec (dotnet/spark#491)
  • CountVectorizer (https://github.com/dotnet/spark/p

Description

Use Case

Looks like it would be handy to add a file_paths_by_partitions method in rust core so language wrappers won't need to manually construct absolute file paths like what we did in python binding: https://github.com/delta-io/delta-rs/pull/221/files

Related Issue(s)

Connect-Databricks.ps1 uses "https://login.microsoftonline.com" as part of the URI to connect. When retrieving a token for a non-AzureCloud tenant (e.g. AzureUSGovernment) the URI root would be different (e.g., "https://login.microsoftonline.us"). As such, cannot use this task to deploy to other tenant types. Would be helpful to be able to specify an Azure Environment and connect to the right e