topic Accessing the S3 Files in Get Started Discussions

Accessing the S3 Files

Databricks3 — Mon, 03 Jul 2023 05:50:15 GMT

I am using the Unity Catalog Cluster. I have a requirement to read the files placed by the source team in a specific location (landing) in S3. I am already using a metastore pointing to a different bucket. Do I need to use an external location pointing to the landing bucket in S3? Additionally, how can I read the data from those files?

Re: Accessing the S3 Files

Anonymous — Wed, 05 Jul 2023 05:30:25 GMT

You have a couple of options to consider:

External Location: You can create an external location in your Unity metastore that points to the landing bucket in S3. This allows Unity to access the files in that location without having to copy or move them to the default location managed by Unity. You can configure the external location using the Unity Catalog's administration tools or by using the Unity SDK/API.

To create the external location, specify the S3 bucket and prefix (folder) where the files are located. Unity will be able to read the data directly from the specified S3 location without any data movement.
Direct Read: Unity also provides the ability to directly read data from files in S3 without the need for an external location. In this approach, you can directly query the files in the S3 landing bucket using SQL or Spark commands. Unity will use its underlying query engine to perform distributed processing and retrieve the data from the S3 files.

To read the data directly from the S3 landing bucket, you can use the Unity Catalog's SQL or Spark interfaces to interact with the data and perform the necessary operations like filtering, aggregating, or joining the datasets.

Re: Accessing the S3 Files

Databricks3 — Wed, 05 Jul 2023 17:43:15 GMT

If you could share an example of reading the file of both the cases it would be really helpful.

Re: Accessing the S3 Files

Anonymous — Wed, 12 Jul 2023 09:54:18 GMT

Hi @Databricks3

Hope you are well. Just wanted to see if you were able to find an answer to your question and would you like to mark an answer as best? It would be really helpful for the other members too.

Cheers!

Re: Accessing the S3 Files

krikotti — Mon, 19 Aug 2024 21:04:19 GMT

did anyone get any solution on this topic? I am also facing the challenges reading the file from s3 using the boto3 with unity enabled cluster, created the s3 external location and granted the enough access. any help on this ?

same path and data accessible using the pyspark without any issues,