如何使用不同的数组长度将 from_dict 提取到 pandas dataframe？

Question

I have a dict, I need to convert to pandas dataframe, dict have arrays, if the arrays are of same length it is working fine, but different length array throwing valueError, second question is I need to access only few key value pairs from the dict

This case working, as expected I get two rows

my_dict = {
          "ColA" : "No",
          "ColB" : [
            {
              "ColB_a" : "2011-10-26T00:00:00Z",
              "ColB_b" : 8.3
            },
            {
              "ColB_a" : "2013-10-26T00:00:00Z",
              "ColB_b" : 5.3
            }
          ],
          "ColC" : "Graduate",
          "ColD" : [
            {
              "ColD_a" : 5436.0,
              "ColD_b" : "RD"
            },
            {
              "ColD_a" : 4658.0,
              "ColD_b" : "DV"
            }
          ],
          "ColE" : "Work"
        }

sa = pd.DataFrame(my_dict)

In this case ColB has only one value

my_dict = {
          "ColA" : "No",
          "ColB" : [
            {
              "ColB_a" : "2011-10-26T00:00:00Z",
              "ColB_b" : 8.3
            }
          ],
          "ColC" : "Graduate",
          "ColD" : [
            {
              "ColD_a" : 5436.0,
              "ColD_b" : "RD"
            },
            {
              "ColD_a" : 4658.0,
              "ColD_b" : "DV"
            }
          ],
          "ColE" : "Work"
        }

sa = pd.DataFrame(my_dict)

so this throws ValueError: arrays must all be same length, How this can be fixed? expected output is

I can do

sa = pd.DataFrame.from_dict(my_dict, orient='index').transpose()

But I have to melt and join again. Second Question, if I need to choose only ColA, ColB from dict to create dataframe, How this to be done?

Answer 1

for your second question, you could select a couple of columns from your dictionary using 'columns' parameter: For example: sa = pd.DataFrame(my_dict, columns = ['ColA', 'ColD'])