Unary and n-ary inclusion dependency discovery in relational databases

Profiling relational data: a survey

“…If the distinct value sets of columns A and B are not available, we can estimate the Jaccard similarity using their MinHash signatures [38]. Conditional functional dependencies [24], [59], CTANE [47], CFUN [42], FACD [91], FastCFD [47] Inclusion dependencies [101], [87], SPIDER [14], ZigZag [102] Conditional inclusion dependencies [61], CINDERELLA [13], PLI [13] Foreign keys [123], [143] Denial constraints FastDC [29] Differential dependencies [128] Sequential dependencies [57] 5 Dependency detection…”

“…De Marchi et al [101] propose a technique that scans the database and builds value indices, which are similar to inverted indices. Table 6 shows excerpts of two relations instances, one with columns A and B and the other with columns C and D, and the corresponding value index.…”

“…Once all unary Inds have been discovered, De Marchi et al [101] give a level-wise algorithm, similar to the TANE algorithm for Fd discovery, which constructs Inds with i columns from those with i − 1 columns and prunes Inds that cannot be true. Additionally, hybrid algorithms have been proposed in [87,102] that combine bottom-up and top-down traversal for additional pruning.…”

“…In the divide step, it splits the input dataset horizontally into partitions and vertically into buckets with the goal to fit each partition into main memory. In the conquer step, Binder then validates the set of all possible inclusion dependency candidates, which are created in the same fashion as in [101], against the partitions. Processing one partition after another, the validation constructs two indexes on each partition, a dense index and an inverted index, and uses them to efficiently prune invalid candidates from the result set.…”

“…Using the notion of removing the fewest tuples so that the remainder satisfies the Ind exactly, we can define the strength or confidence of a partial [96,101]. That is, the confidence is the number of distinct values of X that appear in Y divided by the number of distinct values of X .…”