- From: Farrukh Najmi <Farrukh.Najmi@Sun.COM>
- Date: Wed, 16 Jun 2004 14:13:47 -0400
- To: RDF Data Access Working Group <public-rdf-dawg@w3.org>
Janne Saarela wrote: > > I posed quick question yesterday regarding the programming model > behind this streamable protocol suggestion. > > I am also eager to hear more information about ODBC/JDBC experiences > that someone (Jim?) promised to elaborate by email. I wish to > understand how an ordinary Java/C# programmer could use familiar > constructs such as ResultSets in managing the streaming results. From the perspective of a Java/JDBC programmer ResultSet support if indispensable. For the JDBC ResultSet API document see: http://java.sun.com/j2se/1.4.2/docs/api/java/sql/ResultSet.html http://java.sun.com/j2se/1.4.2/docs/api/index.html (complete JDK API) ResultSet in JDBC allows implementations to stream the results of the query no matter how large. The JDBC implementation is not required to maintain server side state though most do in order to scale. As a late addition to JDBC was to make the ResultSet scrollable so that the client can scroll backward and forward to any position in the result set relative to current position. Also a later addition was the ability to support for Updateable Result set where the client may insert a result anywhere in between the result set and perform an update. The typical application of ResultSet is in Scrollbale UIs where the UI shows a subset of results and displays scroll forward/backward buttons to allow moving to the next and previous window in the result set. A good implementation will do some prefetching of previous and next window to make scrolling over a large result set smoother. I see the proposed streaming requirement as subsuming the proposed Iterative Query requirement and consider it a MUST requirement for a query API. -- Regards, Farrukh
Received on Wednesday, 16 June 2004 14:14:22 UTC