From the graph above, SQL E, which selects data from 2
tables formed the largest bottleneck as it required about 49
seconds to be executed, accounted for almost 20% of the
whole sentence-level novelty detection. This is consistent with
our knowledge as MYSQL is required to scan through 2
different tables which can take up a longer time. The other
peak was the insertion of results, SQL C, after the processing
of one document was completed.