Below is the function I made in order to return last numeric value :

def DerniereValeur(DF):
    for i in range(1,100):
        print(DF[-i])
        if DF[-i].isnumeric():
            ValeurARetourner = DF[-i]
            break
    return ValeurARetourner

Where DF looks like :

Output excepted

DerniereValeur(FirstExample)
> Returns 60.2751
DerniereValeur(SecondExample)
> Returns 18.32

Error Message

AttributeError: 'float' object has no attribute 'isnumeric'

Debug infos

Thrown on DerniereValeur(FirstExample)

line : if DF[-i].isnumeric():

Where DF[-i]=60.2751

DF
Unnamed: 2             100
Unnamed: 3        100.5425
Unnamed: 4       101.01144
Unnamed: 5       101.97366
Unnamed: 6       102.27216
   
Unnamed: 1257     60.97918
Unnamed: 1258    60.568195
Unnamed: 1259    61.285896
Unnamed: 1260     61.92188
Unnamed: 1261     60.62751

CodePudding user response：

Instead of passing columns one by one by the function, you could stack the DataFrame and use groupby.last to get the last numeric values of each column in one go:

out = df.stack().reset_index(level=1).groupby('level_1').last().rename_axis(None).T

Output:

   FirstExample  SecondExample
0       60.2751          18.32

For a single column DataFrame, we could use dropna (to get rid of trailing NaNs) tail (for the last value):

df[['SecondExample']].dropna().tail(1)

Output:

   SecondExample
0          18.32

CodePudding user response：

You could use dropna on the column of interest and then take the last remaining value in the column, checking for IndexError in case there are no non-NaN values:

def DerniereValeur(DF):
    try:
        return DF.dropna().iat[-1]
    except IndexError:
        return np.NaN

df = pd.DataFrame({
    'FirstExample': [28.32, 60.2751],
    'SecondExample': [18.32, np.NaN],
    'ThirdExample': [np.NaN, np.NaN] 
})
DerniereValeur(df['FirstExample'])
DerniereValeur(df['SecondExample'])
DerniereValeur(df['ThirdExample'])

Output:

60.2751
18.32
nan