CODEWITHSUNDEEP
pythonpandas
BlueMango
BlueMango

Reputation: 503

Pandas create new columns according to the values in other columns

Say I have a following DataFrame:

dfff = pd.DataFrame({"name":[["a","b","c","d"]], "value":[[["a","aa"],["b","bb"],["c","cc"],["d","dd"]]]})


    name                         value
[a, b, c, d]    [[a, aa], [b, bb], [c, cc], [d, dd]]

From this I want something like:

   a       b      c        d
[a, aa] [b, bb] [c, cc] [d, dd]

For a single row, this piece of code would work, but it does not if it has multiple rows:

for i in range(len(dfff)):
    for num, val in enumerate(dfff.name[i]):
        dfff[val] = [dfff.value[i][num]]
dfff.drop(["name","value"], axis = 1, inplace = True)

How do I achieve this is pandas? Thanks a lot

Edit: Example for multiple rows table:

    name                         value
0   [a, b, c, d]    [[a, aa], [b, bb], [c, cc], [d, dd]]
1   [c, d, e]       [[c, cc], [d, dd], [e, ee]]

It should become:

   a       b      c        d       e
[a, aa] [b, bb] [c, cc] [d, dd]   None
  None   None   [c, cc] [d, dd]   [e, ee]

Upvotes: 2

Views: 55

Answers (2)

Shubham Sharma
Shubham Sharma

Reputation: 71689

We can zip the columns name and value inside a list comprehension then iterate over the zipped values and for each name-value pair create the corresponding record/dict

pd.DataFrame([dict(zip(*pair)) for pair in zip(df['name'], df['value'])])

         a        b        c        d        e
0  [a, aa]  [b, bb]  [c, cc]  [d, dd]      NaN
1      NaN      NaN  [c, cc]  [d, dd]  [e, ee]

Upvotes: 3

McLean Trott
McLean Trott

Reputation: 37

I think your easiest solution is to just transpose the table.

dfff_transposed = dfff.T # or dfff.transpose()
print(dfff_transposed)

https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.transpose.html

Upvotes: 0

Related Questions