[BUG] Dask print does not print all rows

[goodwanghan]

Minimal Code To Reproduce

%%fsql dask
CREATE [[0,1],[0,2],[1,1],[1,3]] SCHEMA a:int,b:int
PRINT 3 ROWS

Describe the bug
The problem is the head function of dask requires npartitions, and even setting to -1, if a df has 5 records and I want to print 10 rows, I still get the warning which is not nice. So we should revert to itertuples approach.

Environment (please complete the following information):

• Backend: pandas/dask/ray? dask 2021.4.1
• Backend version:
• Python version: 3.7
• OS: linux/windows