CODEWITHSUNDEEP
pythonpandasnumpydata-science
Mihail
Mihail

Reputation: 41

Python DataFrame find values

I've DataFrame like this:

5 1 89107633648 19239.9
6 nan  9348.62
7 nan  3191.83
8 nan  2962.06
9 nan  3737.39
10 2 9156559561 6318.99
11 nan  2138.27
12 nan  1933.3
13 nan  2247.42
14 3 9107812806 25324.19
15 nan  1169.1
16 nan  3750.52
17 nan  3436.4
18 nan  5140.16
19 nan  nan
20 nan  3196.79
21 nan  2985.62
22 nan  2869.01
23 nan  2776.59

I need to get a two-dimensional list grouped by nan field. [[9348.62, 3191.83, 2962.06, 3737.39], [2138.27, 1933.3, 2247.42], ...] How can I do that?

n_rows = res.astype(bool).sum(axis=0)['Unnamed: 5']
rows_3 = [[] for i in range(n_rows)]

for i in range(0, len(res)):
    tmp = 0
    j =0 
    if res.iloc[i]['Unnamed: 5'] is '':
        tmp = res.iloc[i][12]
    else:
       j += 1
    rows_3[j].append(tmp)
    tmp =0
    
print(rows_3)

Upvotes: 0

Views: 106

Answers (1)

sushanth
sushanth

Reputation: 8302

Here is solution, using Series.fillna with ffill then GroupBy followed by list splice.

# renamed columns for convenience
df.columns = ['col1', 'col2', 'col3', 'col4']

print(
    df.assign(col2 = df.col2.fillna(method='ffill'))
        .groupby('col2')['col3'].agg(lambda x: x.tolist()[1:])
        .reset_index(drop=True).tolist()
)

Upvotes: 1

Related Questions