The cursors are designed to move very quickly and efficiently through the document. One can use more cursors but it's not efficient to use a new one for each record.
In the for loop, if possible avoid creating new cursors, create only one at the beginning and use that one for all the records, storing only the bookmarks, both in the map and list. After indexing don't forget to dispose the cursor. For retrieving, you can create only one new cursor and move that one from bookmark to bookmark ( c.toBookmark(bookmark) ) , and finally dispose it. Cezar > -----Original Message----- > From: heikki doeleman [mailto:[EMAIL PROTECTED] > Sent: Tuesday, August 08, 2006 5:50 AM > To: [email protected] > Subject: Using XmlCursor and XmlBookmark > > > [sorry - this time with message subject] > > Hello, > > I'm thinking of using XmlCursor / XmlBookmark to create in andex to a > large > XML document. I'm wondering if I'm taking a right approach, as I found not > many real examples of XmlBookmark usage. Especially I'm wondering whether > the solution sketched below is prone to memory leaks, or otherwise > inefficient. Also, I'm not entirely sure of the exact nature of XmlCursor > vs. XmlBookmark, and it may well be that I'm doing superfluous stuff here; > I > hope you like to take a look. > > Say the XML is like this > > <catalog> > <recordlist> > <record> > <title>short text</title> > <description>much longer text</description> > </record> > . . . . . (many records) > </recordlist> > </catalog> > > My intent is to read this document into memory, then construct an index > such > that each distinct word found in the description elements, acts as a key > that maps to a list of XmlBookmarks representing all records where that > word > occurs. The code is like this: > > CatalogDocument catalogDoc = CatalogDocument.Factory.parse(xmlFile); > Catalog catalog = catalogDoc.getCatalog(); > RecordList recordList = catalog.getRecordList(); > Record[] recordArray = recordList.getRecordArray(); > // the index to construct > Map bookmarxIdx = new HashMap(); > for(int i = 0; i < recordArray.length; i++) { > Record record = recordArray[i]; > XmlCursor recordCursor = record.newCursor(); > String description = record.getDescription(); > StringTokenizer st = new StringTokenizer(description); > while(st.hasMoreTokens()) { > String word = st.nextToken(); > List bookmarksForWord = (List) bookmarxIdx.get(word); > if(bookmarksForWord == null) { > bookmarksForWord = new ArrayList(); > } > XmlCursor.XmlBookmark bookmark = new MyBookmark(word); > recordCursor.setBookmark(bookmark); > bookmarksForWord .add(recordCursor); > bookmarxIdx.put(word, bookmarxValue); > } > } > > // index is created. Now I can quickly retrieve the records that contain > some > // word, for example "hello", doing this: > > List cursorsForSearchterm = (List) bookmarxIdx.get("hello"); > for(Iterator i = cursors.iterator(); i.hasNext();) { > XmlCursor c = (XmlCursor) i.next(); > XmlCursor.XmlBookmark bm = c.getBookmark(MyBookmark.class); > c = bm.createCursor(); > System.out.println("Found record: " + c.getObject().toString() ); > c.dispose(); > } > > The catalog will ermain in application memory during its entire lifetime, > so > I figured that it'd be okay to keep all the XmlCursors I'm storing in the > index -- also during application lifetime and without ever disposing of > them. Then when retrieving records, I retrieve the XmlCursors from the > index, get their bookmark, create a new XmlCursor to position itslef at > that > bookmark, retrieve the record from there, and dispose of this XmlCursor. > > although it works I would like to know whether this is a valid way of > dealing with these XmlCursors and XmlBookmarks, or that I'm way off the > mark > here .. > > thank you so much, and regards > Heikki Doeleman > > > -- > View this message in context: http://www.nabble.com/Using-XmlCursor-and- > XmlBookmark-tf2071921.html#a5704059 > Sent from the Xml Beans - User forum at Nabble.com. > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] _______________________________________________________________________ Notice: This email message, together with any attachments, may contain information of BEA Systems, Inc., its subsidiaries and affiliated entities, that may be confidential, proprietary, copyrighted and/or legally privileged, and is intended solely for the use of the individual or entity named in this message. If you are not the intended recipient, and have received this message in error, please immediately return this by email and then delete it. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
