Repeated rows in a numpy array

Question

I feel like I'm going insane because I can't figure out what feels like should be a simple problem. I want to generate fake data in a numpy array and I can't figure out how to repeat a row of observations. I'd rather generate thousands of rows and I can't figure out how to repeat a row whenever I feel like.

For example, here's my current code:

voters = np.array(
    [
        ['Democrat', 'Republican', 'Third'],
        ['Democrat', 'Republican', 'Third'],
        ['Democrat', 'Republican', 'Third'],
        ['Democrat', 'Republican', 'Third'],
        ['Democrat', 'Republican', 'Third'],
        ['Democrat', 'Third', 'Republican'],
        ['Democrat', 'Third', 'Republican'],
        ['Democrat', 'Third', 'Republican'],
        ['Democrat', 'Third', 'Republican'],
    ]
)

But I just want to be able to condense this. It's obviously not manageable to make large datasets this way!

Thank you

Answer 1

Use np.repeat() :

voters = np.array([['row1', 'row1', 'row1'],
                   ['row2', 'row2', 'row2']])

# We repeat 2 times the first row and 4 times the second row.
np.repeat(voters,[2,4],axis=0)
# voters.repeat([2,4],axis=0) produce the same result.

And we obtain:

array([['row1', 'row1', 'row1'],
       ['row1', 'row1', 'row1'],
       ['row2', 'row2', 'row2'],
       ['row2', 'row2', 'row2'],
       ['row2', 'row2', 'row2'],
       ['row2', 'row2', 'row2']])

Answer 2

You can use this:

np.array([['Democrat', 'Republican', 'Third']]* 10000)

to generate as many duplicate rows as you want.