Fast optimal twig joins

Abstract

In XML search systems twig queries specify predicates on node values and on the structural relationships between nodes, and a key operation is to join individual query node matches into full twig matches. Linear time twig join algorithms exist, but many non-optimal algorithms with better average-case performance have been introduced recently. These use somewhat simpler data structures that are faster in practice, but have exponential worst-case time complexity. In this paper we explore and extend the solution space spanned by previous approaches. We introduce new data structures and improved strategies for filtering out useless data nodes, yielding combinations that are both worst-case optimal and faster in practice. An experimental study shows that our best algorithm outperforms previous approaches by an average factor of three on common benchmarks. On queries with at least one unselective leaf node, our algorithm can be an order of magnitude faster, and it is never more than 20% slower on any tested benchmark query.

A Prime Number Approach to Matching an XML Twig Pattern including Parent-Child Edges

“…The benchmarked datasets used in the experiments and their characteristics are shown in Tables 1 and 2. The selected datasets and benchmark are the most frequent in the literature of XML query processing (Bruno et al, 2002;Lu et al, 2004;Grimsmo et al, 2010;Wu et al, 2012;Li and Wang, 2008;Qin et al, 2007). We generated Random dataset similar to that in (Lu et al, 2004) but we vary the two parameters: depth and fan-out.…”

“…The XML structured queries for evaluation over these dataset were chosen specifically because it is not common for queries, which contain both '//' and '/', to have a significant difference in performance for tightly-structured document such as DBLP and XMark. TreeBank twig queries were obtained from (Lu et al, 2004) and (Grimsmo et al, 2010). Twig pattens over the random data set were also randomly generated.…”

“…Over the past decade, most research in structured XML query processing has emphasized the use of node indexing approaches (Bruno et al, 2002;Lu et al, 2004;Grimsmo et al, 2010;Wu et al, 2012;Li and Wang, 2008;Qin et al, 2007). One of the most important problems in XML query processing is tree pattern matching.…”

“…TwigList (Qin et al, 2007) replaced the complex intermediate storage proposed in Twig 2 Stack with lists (one for every query node) and pointers with simple intervals to capture structural relationships. The authors in (Grimsmo et al, 2010) proposed a new storage scheme, level vector split which splits the list connected to its parent list with P-C edge to a number of levels equals to the depth of the XML tree.…”

Nearest keyword search in XML documents

“…The XPath statement of Q can be expressed as a twig pattern [4,15] shown in Figure 2a, where a single-lined (double-lined) edge represents parent-child (ancestor-descendent) relationship. The goal is to find occurrences of the pattern in the data tree, and for each occurrence, output the value at the position of pnameVal (signified as underlined).…”