CODEWITHSUNDEEP
pythonpython-3.xpandasdataframe
Adam
Adam

Reputation: 95

Looping over specific columns in Pandas

I have this data frame:

data = {'id': [1, 2, 3, 4, 5, 6, 7, 8],
        'stat': ['ordered', 'unconfirmed', 'ordered', 'unknwon', 'ordered', 'unconfirmed', 'ordered', 'back'],
        'date': ['2021', '2022', '2023', '2024', '2025','2026','2021', '1990']    
}
df_data = pd.DataFrame(data)
df_data

I did the following changes to the data frame as follows:

l = [1, 2, 1, 3, 1, 0, 0, 0]
df_d = df_data.join(pd.get_dummies(df_data.date).cumsum(axis=1).mul(lis, axis=0
    ).astype(int))
df_d

I wanted to loop over specific columns and cahnges some values in it like this:

ln = len(df_d.columns)
for i, k in enumerate(df_d.stat):
    if k == 'ordered':
        for column in df_d .columns[-(ln-3):]:
            if df_d.loc[i].at[column] == 1:
                if df_d.columns.get_loc(column) + 3 <= ln:
                    df_d.iloc[i,[df_d.columns.get_loc(column)+2]] = [100]
df_d

The result looks something like this (but it is not complete):

    id  stat           date     1990    2021    2022    2023    2024    2025    2026    2027
0   1   ordered        2021       0       1       1       100    100      1       1      100
1   2   unconfirmed    2022       0       0       2       2       2       2       2       2
2   3   ordered        2023       0       0       0       1       1      100     100      1
3   4   unknwon        2024       0       0       0       0       3       3       3       3
4   5   ordered        2025       0       0       0       0       0       1       1      100
5   6   unconfirmed    2026       0       0       0       0       0       0       0       0
6   7   ordered        2021       0       1       1       1       1       1       1       1
7   8   back           1990       0       0       0       0       0       0       0       0

I want to change all rows that have 'ordered' stat and looping over columns and skip the first 2 values and set all others in that row to 100. But I could not figure out how aviod the skiping step, since it lets some columns behind without changes. After finishing this I waould like to set all values of the last 2 column to 1 in a different step. Any idea on fixing that for-loop problem would be very appreciated.

The expected output:

    id  stat           date     1990    2021    2022    2023    2024    2025    2026    2027
0   1   ordered        2021       0       1       1       100    100    100     100      100
1   2   unconfirmed    2022       0       0       2       2       2       2       2       2
2   3   ordered        2023       0       0       0       1       1      100     100      100
3   4   unknwon        2024       0       0       0       0       3       3       3       3
4   5   ordered        2025       0       0       0       0       0       1       1      100
5   6   unconfirmed    2026       0       0       0       0       0       0       0       0
6   7   ordered        2021       0       1       1      100     100     100     100     100
7   8   back           1990       0       0       0       0       0       0       0       0

Upvotes: 1

Views: 464

Answers (1)

mozway
mozway

Reputation: 262429

IIUC, you can use numpy with numpy.triu to compute a 2D mask and use it to set your 100 values before joining:

df2 = pd.get_dummies(df_data.date).cumsum(axis=1).mul(lis, axis=0).astype(int)

skip = 2 # number of columns to skip
# mask for upper triangle (above +2 diagonal)
m1 = np.triu(np.full(df2.shape, True), k=skip+1)
# mask for rows corresponding to ordered
m2 = df_data[['stat']].eq('ordered').to_numpy()
df2[m1&m2] = 100   # the two masks are combined by broadcasting

df_d = df_data.join(df2)

output:

   id         stat  date  1990  2021  2022  2023  2024  2025  2026  2027
0   1      ordered  2021     0     1     1   100   100   100   100   100
1   2  unconfirmed  2022     0     0     2     2     2     2     2     2
2   3      ordered  2023     0     0     0     1     1   100   100   100
3   4      unknwon  2024     0     0     0     0     3     3     3     3
4   5      ordered  2025     0     0     0     0     0     1     1   100
5   6  unconfirmed  2026     0     0     0     0     0     0     0     0
6   7      ordered  2027     0     0     0     0     0     0     0     0
7   8         back  1990     0     0     0     0     0     0     0     0

Upvotes: 1

Related Questions