Really sparse arrays indexed using hash functions


Hi, I would know if there is differences in terms of performance working on dense or really sparse arrays in SciDB. My problem is the following: I have to create an array where one dimension should be a string. The first option is create a “map” array to list the strings and assign to them an index. The other option is use an hash function that produce a int64 as output to “map” strings and integers (in this case collision won’t be a problem).
In this case the performance will be the same? or the “hashed” version could be slower?