Code improvement: groupby and operation on conditionals

[pythonidae]

I've found a solution for the final block of code, in order to set a 'floor' or minimum to the dataset from the table that was constructed previously.

for idx, row in df.iterrows():
    for analyte in ['compound1', 'compound2', 'compound3']:
        if df.loc[idx, analyte] < float(max_dict[row[0]][analyte]):
            df.loc[idx, analyte] = float(max_dict[row[0]][analyte])

Thank you for your help, perfringo!