Correct syntax to "fit" a classifier in sklearn?

Question

I am trying following simple example:

from sklearn import datasets, svm
iris = datasets.load_iris()
clf = svm.SVC(random_state=0)

For fitting, should I use following statement:

clf = clf.fit(iris.data, iris.target)

Or just:

clf.fit(iris.data, iris.target)

Both above formats have been used in different places, so I am confused.

The first method clf = clf(X,y).fit() seems to be the official version (see here).

I have tried and found that both seem to work, but I may be missing some point here.