I have around 1700 text files stored in s3. I can use EMR notebook with EMR via spark.read.text(s3 path) to read most of them but some of them (around 300) can\'t be read wi