Mastering pandas - Second Edition

Ravel

Ravel helps to flatten the data from multidimensional to 1D:

In [86]: ravel_array = np.arange(0,12).reshape(4, 3)In [87]: ravel_array.ravel()Out[87]: array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11])

The order in which the array is raveled can be set. The order can be "C", "F", "A", or "K". "C" is the default order, where the array gets flattened along the row major, while with "F", flattening occurs along the column major. "A" reads the array elements in a Fortran-like index-based order and "K" reads the elements in the order in which they are stored in memory:

In [88]: ravel_array.ravel(order = "F")Out[88]: array([ 0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11])