Encoding inputs with varying lengths to use in sklearn model

Question

I have the following dataset:

 X = [[{\'feature1\': \'value1\', \'feature2\': \'value2\'},
      {\'feature1\': \'value3\', \'feature2\': \'value4\'},
      {