In Section 3, basic concepts such as query processing and materialized views on XML documents are introduced.
As mentioned in Section 1, one research direction on querying XML documents was to build materialized views on XML documents to reduce the number of visited nodes during tree traversing, thereby leading to faster query processing.
However, a materialized view is the one which is physically stored in the database, in addition to its definition.
In this paper, based on native XML databases, we use materialized views to query required data from an original document.
Then, we build relevant structures such as T-Bitmap and indexes on materialized views to further accelerate query processing.
For the query with an Xpath expression as shown in Figure 4, we can define a materialized view beforehand, which is rooted from node "Books" with an attribute "category" matching with the specified predicate "Technology", as shown in Figure 5.
CREATE MATERIALIZED VIEW mv AS( SELECT extract(sys_nc_rowinfo$j '/root/store/Books[@category="Technology"]') FROM XMLTABLE); Figure 6: Materialized view.
We can build the index tree of each tag using the start values in labeling codes as keys and the well-known B+-tree algorithm, as shown in Figure 10 where the pointers of a leaf node in the tag index tree indicate the positions of corresponding nodes in the materialized view.
Moreover, although we might expect that this ideology would actively work to produce a bipolar distribution of facial hair (and it does to some extent, insofar as women still often remove or dye their facial hair), the fact that these differences are not fully materialized
in twentieth-century western culture may be taken as an index of the diminishing emphasis placed on facial hair.