Find the largest number in column

Question

I'm trying to find the month (column 'Month') that has the largest number (in DepDelay column)

Data

flightID         Month  ArrTime ActualElapsedTime  DepDelay   ArrDelay
BBYYEUVY67527        1   1514.0               58.0       NA      64.0   
MUPXAQFN40227        1     37.0              120.0       13      52.0   
LQLYUIMN79169        1    916.0              166.0       NA     -25.0   
KTAMHIFO10843        1      NaN                NaN        5       NaN   
BOOXJTEY23623        1      NaN                NaN        4       NaN  
BBYYEUVY67527        2   1514.0               58.0       NA      64.0   
MUPXAQFN40227        2     37.0              120.0       NA      52.0   
LQLYUIMN79169        2    916.0              166.0       NA     -25.0   
KTAMHIFO10843        2      NaN                NaN       15       NaN   
BOOXJTEY23623        2      NaN                NaN        4       NaN  

I tried:

data = pd.read_csv('data.csv', sep='\t')

dep_delay = all_data.groupby(["Month"].DepDelay.count().max())

print(dep_delay)

Error:

AttributeError                            Traceback (most recent call last)
<ipython-input-14-2ea6213009d6> in <module>()
----> 1 dep_delay = all_data.groupby(["Month"].DepDelay.count().max())
      2 
      3 print(dep_delay)

AttributeError: 'list' object has no attribute 'DepDelay'

Good output:

Month      DepDelay
    1            22

Answer 1

You need sum instead of count to sum values by group. Here's one way using GroupBy + sum, and then idxmax:

res = df.groupby('Month')['DepDelay'].sum().reset_index()
res = res.loc[[res['DepDelay'].idxmax()]]

print(res)

   Month  DepDelay
0      1      22.0

Alternatively, you can group and sort, then extract the first row:

res = df.groupby('Month')['DepDelay'].sum()\
        .sort_values(ascending=False).head(1)\
        .reset_index()

print(res)

   Month  DepDelay
0      1      22.0

Answer 2

Another approach:

pd.DataFrame(df.loc[df['DepDelay'].idxmax(), ['Month', 'DepDelay']]).T
#  Month DepDelay
#8     2       15

And you can reset index to change 8 to 0:

pd.DataFrame(df.loc[df['DepDelay'].idxmax(), ['Month', 'DepDelay']]).T.reset_index(drop=True)
#  Month DepDelay
#0     2       15