NUMPY
| Functions or Methods | Description |
|---|---|
np.array(object, dtype=None, copy=True, order=’K’, subok=False, ndmin=0) |
Creating a new numpy array, Order: C – row major, F- Column major |
ndarray[3, 4] |
equivalent to ndarray[3][4] |
ndarray.ndim |
Dimension of array, (rows, columns) |
ndarray.itemsize |
Size of each element datatype |
ndarray.dtype |
Data type of elements |
ndarray.size |
Number of elements in the array |
ndarray.shape |
(rows, columns) |
ndarray.reshape(row, column) |
Change the current dimension to given dimensions |
ndarray.sum(axis=0) |
Total sum if axis not given else 0 – column wise, 1 – row wise |
ndarray.ravel() |
Converting ndarray to one dimensional array |
ndarray.dot(ndarray2) |
Dot product of two matrix |
ndarray.T |
Transpose of ndarray matrix |
ndarray.max() / min() |
return max/min value of array |
ndarray.argmax() / argmin() |
return index of max / min value in array |
arr_copy = ndarray.copy() |
copy array to another variable # arr2 = ndarray[0:n] => it doesnt copy part of ndarray and put in arr2, but just reference, so any change in arr2 will reflect in ndarray |
np.linespace(start, end, values) |
Equally separated numbers from start to end, values is no of elements required |
np.logspace(start, end, values) |
Equally separated log(numbers) from start to end |
np.arange(start, stop, steps) |
like range method, ‘steps’ separated values from start to stop |
np.zeros((rows, columns)) |
Zero matrix, all elements are 0, inside is a tuple |
np.ones((rows, columns)) |
Ones matrix (all elements are one) |
np.eye(rows, column) |
Identity matrix, has diagonal of 1 |
np.empty((rows, column)) |
Matrix with random numbers |
np.std(ndarray) |
Standard deviation of each element of ndarray |
np.nditer(ndarray) |
Iterate through each element of ndarray |
np.sort(ndarray, axis=-1, kind=’quicksort’, order=None) |
Sorting numpy array |
np.sqrt(ndarray)np.exp(ndarray)np.max(ndarray)np.sin(ndarray) |
basic operations |
np.random.rand(row, col) |
2D matrics of row*col random numbers from 0 to 1, can pass single arg for 1D array |
np.random.randn(2)# can be randn(row, col) |
2 numbers equally distant from 0 => (0.1748905, -0.917962014) |