The paper 'Piggyback Proﬁling: Enhancing Query Results with Metadata' by Claudia Exeler, Maria Graber, Tino Junge, Stefan Ramson (HPI), Cathleen Ramson, Fabian Tschirschnitz, and Felix Naumann (HPI) was accepted at LWDA conference 2018 in Mannheim.
For a short Abstract see below:
SQL-based data exploration is tedious. Any given query might be followed up by another query simply to count the number of distinct values of an interesting column or to ﬁnd out its range of values. Besides the actual query output, each result also embodies various metadata, which are not visible to the user and not (yet) determined by the DBMS. These metadata can be useful to understand the data, assess its quality, or spark interesting insights.
Our approach piggybacks metadata calculation on the usual query pro-cessing with minimal overhead, by making use of speciﬁc properties of the query plan nodes. We describe our extension of an RDBMS and show that its runtime overhead is usually less than 10%.