2 Answers
- Newest
- Most votes
- Most comments
0
This SageMaker Canvas error is typically related to IAM permissions or configuration issues. You could check the most common causes and solutions for your S3 import problem:
Cause & Solution:
Ensure your role has permissions to;
- "AmazonSageMakerCanvasFullAccess" policy
- S3 read/list permissions for your bucket & it's contents
- Ensure Canvas and S3 bucket are in the same region
- Check bucket policies don't block SageMaker access
- Ensure your tabular dataset is in supported format e.g. CSV, Apache Parquet, Apache ORC, JSON
Please refer: Grant Permissions for Cross-Account S3 Storage
- Your SageMaker execution role might have trust relationships with other AWS services, you can check if your role trusts services beyond SageMaker. Edit the trust policy to only include SageMaker:
{
"Version":"2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Principal": {
"Service": [
"sagemaker.amazonaws.com"
]
},
"Action": "sts:AssumeRole"
}
]
}
Please refer: SageMaker Canvas Limitations and Troubleshooting
You might wanna check this previous reported issue as well;
answered 9 months ago
0
Canvas only supports tabular data in specific formats:
Supported
- .csv
- .parquet
- .tsv
- .xlsx (Excel)
- .json (newline-delimited for tabular)
Ensure
- Try downloading the file locally and opening it
- Ensure the first row is a clean header row
- No merged cells (for Excel)
Relevant content
asked 3 years ago
asked 2 years ago
