I was reading single pass in memory indexing and had few doubts.
Why is the time complexity of SPIMI O(T) where T are the token. We know that before writing blocks to Disk, we have to sort dictionary terms. So, why don't we consider log(T) factor of it?
Also, if there is some data in memory then while performing sorting can we sort it in log(n) time if all of the main memory is exhausted?