REL: A structural query language for processing massive line-based records

摘要

Many modern programs still use flat files to store application data because of their simplicity and portability. These files contain massive line-based records, on which common actions performed involve filtering, mapping and basic set operations. Rather than having each program to implement them on its own, we present REL, a structural query language that handles all these complexity behind and provides an elegant query interface. We also implement REL as a UNIX utility to translate query expressions into UNIX shell scripts that do the actual record processing. Finally we evaluate its performance through experiments.