In general, if you are having performance issues with highlighting, the first thing to do is double check what the bottleneck is: is it accessing the text to by highlighted, or is it running the highlighter?
you suggested earlier in the thread that the problem was with accessing the text... : >>> Now we keep this big text field on disk (in a file), and feed it to : >>> the highlighter. Unfortunately the highlighter has to read the file, : >>> parse it, etc... It's slooow, sometimes over a second on a large the first question to ask is: are you storing just the text that you want to highlight, or is it in a higher level format -- i ask because you use the expression "parse it" which suggests to me that perhaps you are storing raw HTML or RDF odcuments or something like that ... things might be faster for you if you store only the extracted text. second: for many applications, only a small percentage of the documents every wind up getting surfaced via search, and more importantly: a small percentage tend to be surfaced a high percentage of hte time ... perhaps you could add an in memory cache of the most frequently highlighted documents (without hte highlighting being applied, so cache hits can occur on popular documents even if the search words are different) if you find that accessing the text isn't the bottleneck, and the acctual highlighting call is what's slow, then perhaps the issue is that the fragmenter is working too hard to come up with good snippets, and you could write your own that does leswork but takes less time? -Hoss --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
