Hi I am looking into how best to support updating of indexes incrementaly with support for replacing old documents. As I see it from what I have read and tried. There are two ways of attempting this.
Option A 1)Open only a reader or search. From a list of documents to be updated, find all those already indexed, using an IndexReader (or Searcher), and get the document IDs. 2)For each ID delete it using an IndexReader. 3)Then close the reader, then open a writer, and index all the new documents. 4)Close the writer, and reopen all readers and searchers where necessary. Option B 1)Open only a writer and index all the documents, keeping track of any occurenct of an existing document. 2)Close the writer, and open a reader. Delete all updated documents. Continue to use the reader or searchers, reopenning to flush to disk if necessary. Would these be the two appropriate solutions, or are their better ways to do it (without creating a second index and then switching between them). In particular can you avoid either having duplicate records from option A, or missing records from option B whilst in the middle of the process. Thanks Tony --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
