Hello everyone. It seems pyspark dataframe read is broken for reading multiple files.
sql.read.json( "file1,file2") fails with java.io.IOException: No input paths specified in job. This used to work in spark 1.4 and also still work with sc.textFile Blaž