Re: Count Sum of Term Occurrences?

[Ahmet Arslan]

> How can I count the total number of a
> specific terms occurrences?
> 
>  
> 
> How can you get the total number of occurrences of a term
> across all
> documents (e.g. Sum of the number of occurrences of a
> specific term in each
> doc)? 
> 
>  
> 
> For example, I have 3 documents, document #1 has "The green
> bird is flying"
> document #2 has "The green Car has a green driver", and
> document #3 has "I
> just love the color green, oh green, such a nice green, I
> wish I were
> green". 
> 
>  
> 
> I know the Terms component will give me the number of
> documents which have
> the word green (in my example '3') but I want the sum
> occurrences (in  my
> example '7').

So you want collection frequency instead of document frequency. You can modify 
TermsComponent.java to do that by appending this code snippet after the line 
'int docFreq = termEnum.docFreq();' :

TermDocs termDocs = rb.req.getSearcher().getReader().termDocs(theTerm);
int collectionFreq = 0;
while(termDocs.next())
collectionFreq += termDocs.freq();
docFreq = collectionFreq;