Difference Between numpy.genfromtxt and numpy.loadtxt, and Unpack

Question

I am curious to know the difference between the two functions alluded to in the title of this thread. From the website containing the documentation, it says, "numpy.loadtxt [is] [an] equivalent function when no data is missing." What exactly is meant by this? Does this mean, for instance, if I have a csv file that has a blank column between two columns containing data, I should not numpy.loadtxt?

Also, what does this mean,

"unpack : bool, optional If True, the returned array is transposed, so that arguments may be unpacked using x, y, z = loadtxt(...)" 

I am not quite certain as to what this means.

I'd appreciate your help, thank you!

Answer 1

You are correct. Using np.genfromtxt gives you some options like the parameters missing_values, filling_values that can help you dealing with an incomplete csv. Example:

1,2,,,5 6,,8,, 11,,,, 

Could be read with:

filling_values = (111, 222, 333, 444, 555) # one for each column np.genfromtxt(filename, delimiter=',', filling_values=filling_values)  #array([[   1.,    2.,  333.,  444.,    5.], #       [   6.,  222.,    8.,  444.,  555.], #       [  11.,  222.,  333.,  444.,  555.]]) 

The parameter unpack is useful when you want to put each column of the text file in a different variable. Example, you have the text file with columns x, y, z, then:

x, y, z = np.loadtxt(filename, unpack=True) 

Note that this works the same as

x, y, z = np.loadtxt(filename).T 

By default iterating over a 2-D array means iterating over the lines, that's why you have to transpose or use unpack=True in this example.