Argthtjtr

If I have a 3xn numpy array full of integers, how could I easily add those integers into a DataFrame like so:

input:

a = np.array(array([[2,10, 8],[2,9,4],[8,2,2],[8,9,10],[2,3,8]])

output:

Create empty DataFrame at first with columns being each unique integer from the numpy array.

In each row of this new DataFrame, I need to have each row in the numpy array appended to the unique columns. Each new integer should turn into a column separately containing the whole row.

Should look something like this before it gets to the [2,3,8] in the array:

Column:    2       4       8        9       10

R      |[2,9,4]|[8,2,2]|[2,9,4] |[8,2,2]| [2,9,4] |

O       [8,2,2]         [2,3,8]  [2,3,8]  [2,3,8]

W:      [8,9,10]        

Since the last entry in the numpy array has a 3 in it, and 3 is not yet in the column list, I'd like a new column to be made for it.

For example if the next item in the array after [2,3,8] is [1,89,2] then the DataFrame should now look like:

Column:    2       4       8        9       10         3        1     89

R      |[2,9,4]|[8,2,2]|[2,9,4] |[8,2,2]| [2,9,4] | [1,89,2] |     |     |

O       [8,2,2]         [2,3,8]  [2,3,8]  [2,3,8]

W:      [8,9,10]        [1,89,2]

        [1,89,2]        [8,9,10]

1 and 89 are now created as rows awaiting the next item in the numpy array.

Then columns 1, 89 and 2 would contain the next item and so on. Hopefully this is more clear.

Technically I don't care how the data is stored, I assumed a dictionary at first but the DataFrame makes more sense when I look at it now. If there is a better way with a list, a dictionary or some other function I'm not aware of, please let me know if it makes sense now.

2nd Edit:

Sorry for the confusion guys.

dict1={'2':[[2,9,4],[8,2,2]],'4': [[8,2, 2],[8,9,10]],'8':[[2,9,4],[1,89,2]],'9':[[8,2,2],[2,3,8]],'10': [[2,9,4],[2,3,8]],'3': [[1,89,2],[2,3,8]]}



df3=pd.DataFrame(dict1)

I think this will help you