Consider empty columns when loading CSV with Python

I'm programming in Python and reading from a CSV template with Numpy like this:

csv = np.genfromtxt('MMRExport.csv',delimiter=",", dtype=str)

But inside the CSV I have rows of empty data and it doesn't take them into account, for example the first row has:

"A","B","C","","D","E",""

Then the columns that have empty fields are not considered or it throws me an error in the cases that following rows have data in that location. For example, if row 2 is the same shape as row 1 there is no problem, but if another row was of the form:

"1","","2","3","4","5","6"

It throws me an error since the number of columns are different. What I want is to read all the columns even the empty ones. How could I do that?

The error it throws me is this:

    Traceback (most recent call last):
      File "/Users/JPontigo/Downloads/Tarea.py", line 17, in <module>
      csv = np.genfromtxt('MMRExport.csv',delimiter=",", dtype=str)
      File "/Users/JPontigo/anaconda/lib/python3.5/site-     packages/numpy/lib/npyio.py", line 1769, in genfromtxt
      raise ValueError(errmsg)
      ValueError: Some errors were detected !
           Line #2 (got 27 columns instead of 5)
           Line #3 (got 27 columns instead of 5)
           Line #4 (got 27 columns instead of 5)
           Line #5 (got 27 columns instead of 5)
           Line #6 (got 2 columns instead of 5)
           Line #7 (got 27 columns instead of 5)
           Line #8 (got 27 columns instead of 5)
           Line #9 (got 27 columns instead of 5)
           Line #10 (got 27 columns instead of 5)

The file is this:

https://drive.google.com/file/d/0B11sJdX_AaJBWGNabERCMzhiYVk/view?usp=sharing