It's for a cursor not a table, parameters will remain while the cursor exists. So there is some sense in minimizing the size, because I need do rebuild the whole index on every refresh

>even worse: you wouldhave to reindex the whole file each time one of the parameters changes. The beauty of setting the UDF as index is that it only recalculates changed fields index for that particular row - so do NOT make the function dependant on a single field! Pre-set numbers give you HUGE advantage here ;-))
>
>>MIN (240,max(count(sections))*tnLength)
>>
>>but you are right, 240 is good enough and possibly it is faster to have the long index then to calc the len and min over and over