How to save to_csv with python

Question

I'd like to have these items saved to_csv as one document

df = pd.DataFrame({'STREAM':['EAGLE','HAWK','HAWK','HAWK','EAGLE','HAWK','EAGLE'],'MAT':['A','D','F','D','C','C','E'],'KIS':['B','D','E','D','A','C','D'],'GEO':['B','C','E','E','F','A','B'],'BST':['C','E','D','D','B','F','C']})

columns = ["A",'A-',"B","C","D","E", "F"]
a = df.melt(id_vars=['STREAM'], value_vars=['MAT','KIS','BST','GEO']).pivot_table(index='STREAM', columns='value', values='variable', 
 aggfunc='count', fill_value=0, margins=True, margins_name='TOT').rename_axis(None)

print('SHOW FIRST TEXT HERE')
print(a)
print()
print('SHOW SECOND TEXT HERE')
print(df)

Such that my outcome would be something like this

       A  B  C  D  E  F  TOT
EAGLE  2  4  3  1  1  1   12
HAWK   1  0  3  6  4  2   16
TOT    3  4  6  7  5  3   28

SHOW SECOND TEXT HERE
STREAM MAT KIS GEO BST
EAGLE   A   B   B   C
HAWK    D   D   C   E
HAWK    F   E   E   D
HAWK    D   D   E   D
EAGLE   C   A   F   B
HAWK    C   C   A   F
EAGLE   E   D   B   C

Anybody with leads to this to assist

Answer 1

You can concatenate the two dataframes:

pd.concat([a, df], axis=0).to_csv('concatenated_dataframes.csv')

Although the differing shapes will mean you have NaNs filling the 'gaps'

Answer 2

Saving to csv usually means saving one table with the same headers. if you want to maintain the shapes of your dataframes, saving to MS Excel is a good option, since you can save them in differents tabs or spreadsheets .

If you like, you could use pickle . This would allow you to save each of your dataframes to the pickle, perhaps in a dictionary.

import pickle
dict = ('first': df, 'second': a)

pickle.dump(dict, open( "yourfile.pickle", "wb" ))

Then to open back up. pickle.load opens the file. In this case you stored two dataframes in one dictionary, so you will get a dictionary with two dataframes when you open/load the files from pickle.

new_dict = pickle.load( open( "yourfile.pickle", "rb" ) )

# To then get your dataframes, 
df = new_dict['first'] 
a = new_dict['second']

HDF5 is probably more complicated than you need.