3

Maybe a very vague question, but digging the links on numpy did not help me.

I need to do a similarity matrix calculation with following hierarchial clustering for binary array that look like this 

name    val1    val2    val3    val4    val5
comp1   0   0   1   0   1
comp2   1   0   0   0   0
comp3   0   0   1   0   0
comp4   1   1   0   0   0
comp5   0   0   1   0   0

I don't understand the concept of row names in numpy. I can read the file like this

test = np.genfromtxt('test.b', delimiter='\t', names = True, dtype = None)
print type(test[0])
numpy.void
print test[0]
('comp1',0, 0, 1, 0, 1)

But how to take into account the row names (this info is very important)? Is it possible?

I suppose that the void is not a correct way of storing a binary array for further similarity matrix calculation?

Improve this question

1 Answer 1

Reset to default
6

Numpy doesn't really support row names. It does support column names, through structured arrays. You could use something like dtype=[('name', object), ('val1', int), ...]. That could also be automated by reading the first line of the file, maybe.

What genfromtxt is giving you is simply an array of type object, where one column happens to contain strings and the others happen to contain integers – but all of them are stored inefficiently as Python objects, rather than in efficient formats.

You may be interested in pandas, which extends numpy matrices with support for labeled rows (among many other things). pandas.read_table will handle your file nicely.

Improve this answer
Sign up to request clarification or add additional context in comments.

Comments

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.