Feature group access issue from kubeflow using python api

[Sangeeth]

We are trying to integrate Hopsworks with Kubeflow and access the feature group from kubeflow jupyter notebook using Python API, and we are able to access a small feature group with less number of elements, but failing to access a bigger feature group.

We are trying to access only small amount of data from the specific feature group, but it is failing and the kernel got disconnected after ample amount of waiting.
the code is given below

import hsfs

connection = hsfs.connection(host=“uuid”,
project=“project_name”,
engine=“hive”,
api_key_value=“key-value”)

fs = connection.get_feature_store()
target_fg = fs.get_feature_group(“feature_name”, version=1)
target_fg.show(5)

[moritzmeister]

@Sangeeth

We are trying to integrate Hopsworks with Kubeflow and access the feature group from kubeflow jupyter notebook using Python API, and we are able to access a small feature group with less number of elements, but failing to access a bigger feature group.

Hi Sangeeth,

when you say “elements” do you mean rows or number of features?
In general, if it works for a smaller number of rows, a dying Kernel is often a sign of too little memory, because Python won’t throw an OOM Error in this case. Can you make sure you have enough resources available?