I've found the answer. I am using the CustomScoreQuery and the doc id
inconsistency is because of the segment based IndexReader concept.
On Fri, Jul 27, 2012 at 3:56 PM, Phanindra R <phani...@gmail.com> wrote:
>
> Hi,
> This is the first time I am encountering this kind of scenario. Lack
> of knowledge might be the reason for my surprise.
>
> - I am using myIndexWriter.explain(Query q, int docId).
>
> - For the same query, the explanation for most of the docs is exactly what
> it's supposed to be. However, for some docId, the explanation is showing a
> different doc id. That doc id doesn't even have the passed in query term.
> The following is the output.
>
> Explaining: Doc id: *73714*, Query: MY_FIELD:bike
>
> 60.828213 = (MATCH) fieldWeight(MY_FIELD:bike in *22973*), product of:
> 1.7320508 = tf(termFreq(MY_FIELD:bike)=3)
> 70.23837 = idf(docFreq=272, maxDocs=438145)
> 0.5 = fieldNorm(field=MY_FIELD, doc=*22973*)
>
> You may notice that the underlined doc ids are different from each other.
> The "same" code is working perfectly fine for other document ids.
>
> Here is the code snippet, just in case.
>
> private void explainQuery(Query q, int doc) throws IOException {
> System.out.println("Explaining: " + doc + ", Query: " + q );
> Explanation exp = is.explain(q, doc);
> System.out.println(exp);
> }
>
> I am not able to figure out the reason behind it.
>
> Thanks,
> Phani
>
