EMR - Expected schema-specific part at index : s3:

When reading data files from S3 in AWS EMR Spark or when submitting spark-submit command, the following exception can popup:

java.net.URISyntaxException: Expected schema-specific part at index * : s3:
at org.apache.hadoop.fs.Path.initialize(Path.java:263)

Debug this issue

The error message itself already provides hint about where the issue occurs: the URI is not correct or invalid format.

Thus to fix this issue, please check anything related to URL in your spark-submit command or Spark scripts.

For instance, the following command has spaces in the URL:

spark-submit --py-files "s3://mybucket/ spark/package.zip" ...

To resolve this issue, you can remove the space in the S3 path. 

References

Uniform Resource Identifier - Wikipedia