This attribute name is a bit misleading. Actually it means the fetch size for the result set. The fetch size describes how much datasets will be read at once. The default value differs a bit and is mostly 64 or 100.
In case of reading a lot of datasets it is recommended to use a fetch size about 10000.
There is a individual maximum which value increase the performance. Please keep in mind all datasets inside the fetch size window will be kept in memory.
