i'm just wondering if there's someone here who have tried to check if there's a performance difference between using long and float/scaled_float in elasticsearch specially when using sum aggregation?
for the long, the number will be multiplied first by 100 before storing it in elasticsearch.
we are not trying to solve the precision error on float but rather we are trying to check if there will be a performance improvement when aggregating if we use long instead of float/double
Thank you