I have an issue: I write sorted (by the specific key e.g. timestamp) data to HDFS into multiple files (each file ~70GB and data is sorted). I want to read this files by Flink. But I have to be sure in ordering (sorting) of data (across the files).
My question is: how does Flink read files from HDFS? Does it read file by file or read it randomly by blocks of each file?