I came across this code in Python:
processed = data[:,feature_locs]
I tried testing this sort of code out with this
ha = (3, 5, 7)
print ha
data = [1, 2, 3, 4, 5]
print data[:, ha]
but I get a TypeError.
What am I doing wrong, and what does the above syntax mean?
This is a custom slice argument supported by the numpy module for multidimensional arrays.
>>> import numpy
>>> a = numpy.random.random((2,3))
>>> a
array([[ 0.01211291, 0.06738324, 0.11690497],
[ 0.86175703, 0.21903569, 0.49506358]])
>>> a[:,1]
array([ 0.06738324, 0.21903569])
See the numpy documentation for more details http://docs.scipy.org/doc/numpy/reference/arrays.indexing.html
when you are working with a 2 dimensional array (or list of lists) using this syntax addresses the whole first dimension and the range specified in the second argument. as such -
s = [11, 12, 13, 14]
s2 = [21, 22, 23, 24]
s3 = [s, s2] # List of lists
s4 = array(s3) # 2-dimensional array
print s4[:,1:3]
-> [[12, 13],
[22, 23]]
as you can see, this runs over the first dimension entirely and prints the second and third element of the second dimensions
EDIT: as mentioned by @tobias_k in the comments - the array() function is provided by the numpy package so you need to add a proper import for it. this syntax is provided by the package for 2-dimensional arrays and not for ordinary list-of-lists.
array is a function from the numpy package. It does not work with "ordinary" lists of lists.
numpy