I have found that mminfo takes an extremely long time
to return when a document has many instances of
a search term. Is there some known inefficiency in this
case?
What options are you using to mminfo, and what is many? Since Texis does need to build up a large string, you may want to limit the number of hits it will enumerate.
I'm using , 0, 0, 3) as the parameters to mminfo.
One example had 17,000 hits in a document and it took hours to return.
We don't have the option of limiting the number. We need to get all of them.
I can't imagine why mminfo would be that slow. What's the query? What non-default apicp settings? How big is the document being searched?
Aren't you folks programming in C? If so, you might be better off using the mm api directly rather than using mminfo() then having to parse out the result. See openmmapi(), setmmapi(), getmmapi(), infommapi(). http://www.thunderstone.com/site/texisman/node231.html
