通过根据另一个数据帧上的条件选择pandas数据框列来创建numpy数组

Question

I am newbie to Pandas and I am trying hours to solve my problem but I haven't manage to solved it. I have seen several stackoverflow conversations but none it is similar to my problem.

I have two data frames, the training with shape (256, 800) and the labels (1, 800), each label correspond to one digit 0-9 and I want to create 10 numpy arrays where each contains only the training data of the i digit.

If the first training example (1*256) has label 0 to assign the column (1*256) of the training to the first numpy array and so on. ie iterating one dataframe assign to a numpy array based on condition to another dataframe. If for example I have 89 "zeros" the 'zeros' numpy array will by 256*89..

Thanks in advance

Answer 1

This should work: dict is the dictionary and keys refer to the labels. I am assuming that labels is a one row multiple columns dataframe, that is why I transposed it.

dict = {0:[], 1:[], 2:[], 3:[], 4:[], 5:[], 6:[], 7:[], 8:[], 9:[]}
for i in range(df.shape[0]):
    dict[labels.T.[i]].append(df.iloc[i, :])