pandas, grouping and aggregating

Question

I have grouped a dataframe:

rwp_initial.df.loc[rwp_initial.df.sample_name=='sma_initial'].groupby(by=['sample_name','pH','salt','column'])['concentration'].plot(marker = 'o', rot=30)

and get following output:

sample_name  pH   salt  column
sma_initial  5.7  50    5         Axes(0.125,0.125;0.775x0.775)
                        6         Axes(0.125,0.125;0.775x0.775)
                  100   7         Axes(0.125,0.125;0.775x0.775)
                        8         Axes(0.125,0.125;0.775x0.775)
                  200   9         Axes(0.125,0.125;0.775x0.775)
                        10        Axes(0.125,0.125;0.775x0.775)
                  400   11        Axes(0.125,0.125;0.775x0.775)
                        12        Axes(0.125,0.125;0.775x0.775)

i would like to take the mean within each pH and salt concentration. The columns are just the same sample measured two times. If i use aggregate(np.mean) the average of all datapoints of one column is calculated.

This figure maybe highlights the data points i would like to take the average of ( i would like to average along the rows):

rwp_initial.df.loc[rwp_initial.df.sample_name=='sma_initial'].groupby(by=['sample_name','pH','salt'])['concentration'].plot(marker = 'o', rot=30)

Answer 1

OK, i found the answer:

grp_initial = rwp_initial.df.loc[rwp_initial.df.sample_name=='sma_initial'].groupby(by=['sample_name','pH','salt']).concentration

for grp, val in grp_initial:
    print(val.groupby(level='row').aggregate(np.mean))

works