Sagemaker Canvas- Data set load error

When encountering a dataset load error in SageMaker Canvas after data transformation, there are several potential issues that could be causing the failure without a specific error code.

First, SageMaker Canvas performs data validation before building models to identify issues that might cause build failures. Even though your initial validation was successful, there might be underlying data quality issues that weren't caught initially but emerged during the model building process.

Common issues that could cause your dataset to fail include:

1. Missing values in your target column
2. Too many unique labels in your target column for a multi-class model
3. Non-numeric values in your target column if the model requires numeric inputs
4. Column names containing double underscores
5. Incomplete rows in your dataset
6. Schema inconsistencies after transformation

To troubleshoot this issue:

1. Try validating your data again by choosing "Validate data" on the Build page to see if any issues are now detected
2. Check your target column for missing values or too many unique classes
3. Verify that your column names don't contain double underscores
4. Ensure all rows in your dataset are complete after transformation

If the issue persists, you could try:

1. Making changes to your data flow in Data Wrangler to address potential issues
2. Creating a new version of your dataset with modifications
3. Checking if your dataset size (50k rows with 11 columns) exceeds any service limits

You can also edit your data flow steps without deleting the transforms already applied. This allows you to make changes to your data source or dataset configuration while keeping the steps of your data flow intact.
Sources
Data validation - Amazon SageMaker
Update a dataset - Amazon SageMaker
Edit data flow steps - Amazon SageMaker

Removing "Column names containing double underscores" resolved the issue. Thanks