result is the co-opted name of a current G graph store in the current database, only G machines write to it. result is a standard G graph at all times.
SQL programmers can load up result with very complex search graphs, then run from result, as if it were a procedure, even with a proprietary base of SQL procedures. The all the data bases being searched need not even be in G format, the onty requirements is that result be writeable and always in G format. Otherwise, installed G procedures can be unrelated to G graph, tables without triples.
The advantages here is that existing tables in SQL land traverse quite nicely, just add an associated schema subgraph that leads to a config file and jump to:
G1 = a:b:config.(link1,link2,link3,link4)
Then @(*,G1) can load up the operators for the table format and follow the links in succession. Run G with schema only, then descend into proprietary table formats. The schema network, itself a G graph, is generally constructed by other convolutions leaving patterns to data, in G World. Dedicated searchers leave schema graphs rather than bookmarks.
No comments:
Post a Comment