topic Re: File not found error when trying to read json file from aws s3 using with open. in Get Started Discussions

File not found error when trying to read json file from aws s3 using with open.

Nagarathna — Mon, 15 Apr 2024 07:36:08 GMT

I am trying to reading json from aws s3 using with open in databricks notebook using shared cluster.

Error message:No such file or directory:'/dbfs/mnt/datalake/input_json_schema.json'

In single instance cluster the above error is not found.

Re: File not found error when trying to read json file from aws s3 using with open.

artsheiko — Tue, 16 Apr 2024 12:05:16 GMT

Hey,

Please consider to use Unity Catalog with Volumes.

You'll find a quickstart notebook here : https://docs.databricks.com/en/connect/unity-catalog/volumes.html#tutorial-unity-catalog-volumes-notebook

Hope it helps,

Best,

Re: File not found error when trying to read json file from aws s3 using with open.

Nagarathna — Wed, 17 Apr 2024 05:12:46 GMT

Hey,

Thanks for suggesting this approach.

But I want to know why the json file cannot be read from AWS S3 bucket using "with open" in python with shared instance mode cluster.

The code works perfectly fine if I'm using a single instance mode cluster.

code:

with open("/dbfs/mnt/datalake/input.json","r") as f:

data = f.read()

Re: File not found error when trying to read json file from aws s3 using with open.

NandiniN — Wed, 01 May 2024 10:07:24 GMT

Hi @Nagarathna ,

I just tried it on a shared cluster and did not face any issue. What is the exact error that you are facing? Complete stacktrace might help. Just to confirm are you accessing the "/dbfs/mnt/datalake/input.json" from the same workspace when trying out the cluster modes?

Thanks!