Tweet
Using the xml output option, it appears that the number of search results returned doesn't match the number of results found. If there are more than 1000 results found, it only returns a subset of the results.
-
-
That's normal and is not limited to the XML output. If you try to browse beyond the 1000th result in large result sets, you should get a message: "You have requested more results than served per query. Please try again with a more precise query."
If you think about it, it's not ever really meaningful to look at results that far down the list (the 1001th most relevant search result?). Google has the same restriction if you try to browse or request more than 1000 results, eg.:
http://www.google.com/search?q=test&start=1000 -
Thanks Ray, that makes total sense. But I don't see any message. The search results listing just stops and looks truncated. Is that supposed to be a message generated by the ZoomIndexer?
In any event we'll build something to handle it. Thanks again.Comment
-
No, the message should only be displayed by the CGI when the user requests more than 1000 results. What request are you sending to the CGI - can you give us an example?Comment
-
We use Zoom to index a huge pdf file library. So if I search for something like "total" I get 1543 results. The cgi doesn't come back with any message, just displays the results as normal. But it's ok we've accounted for it ourselves since we're using the xml output. We just put our own message up if there are > 1000 results.Comment
-
I think that might be the best way to handle it (displaying your own message). The reason is that it's not really an "error" in that case, and the abovementioned message (which is marked as <zoom:error>) would be implying that something has gone wrong (which it hasn't). The XML output has to be OpenSearch and RSS compliant so we have to consider how other uses may interpret this message. Nonetheless, it is something to consider should we need to make changes to the XML format in the future.Comment
Comment