In this notebook, we'll compare MultiIndex and String representations of measurements.



In [25]:

    
from __future__ import print_function, division
import pandas as pd
from numpy.random import randn
from StringIO import StringIO

Creating and reading in data

Create DataFrames from random data

MultiIndex



In [2]:

    
columns_tuples = [('power', 'active'),
                  ('power', 'apparent'),
                  ('energy', 'active'),
                  ('energy', 'apparent'),
                  ('voltage', '')]

index = pd.MultiIndex.from_tuples(columns_tuples, names=['physical quantity', 'type']) 

pd.DataFrame(randn(2,5), columns=index)









    Out[2]:






  
    
      physical quantity
      power
      energy
      voltage
    
    
      type
      active
      apparent
      active
      apparent
      
    
  
  
    
      0
       0.698756
       1.366798
       2.320818
      -1.592383
       0.155624
    
    
      1
      -0.817807
       2.438316
       1.541125
      -0.393314
       0.693330

String



In [3]:

    
columns_strings = [str(p) + (',' + str(t) if t else '') for p,t in columns_tuples]

pd.DataFrame(randn(2,5), columns=columns_strings)









    Out[3]:






  
    
      
      power,active
      power,apparent
      energy,active
      energy,apparent
      voltage
    
  
  
    
      0
      -0.979919
      -0.431717
      -0.073521
       0.731000
       0.109189
    
    
      1
       0.494156
       0.410478
       0.820294
      -0.812069
       0.256516

Read from CSV



In [4]:

    
CSV = """0 1 2 3 4 5
10 11 12 13 14 15
20 21 22 23 24 25
"""

MultiIndex



In [5]:

    
multiindex_df = pd.read_csv(StringIO(CSV), sep=" ", names=columns_tuples)
multiindex_df.columns.set_names(['physical quantity', 'type'], inplace=True)
multiindex_df









    Out[5]:






  
    
      physical quantity
      power
      energy
      voltage
    
    
      type
      active
      apparent
      active
      apparent
      
    
  
  
    
      0 
        1
        2
        3
        4
        5
    
    
      10
       11
       12
       13
       14
       15
    
    
      20
       21
       22
       23
       24
       25

String



In [6]:

    
string_df = pd.read_csv(StringIO(CSV), sep=" ", names=columns_strings)
string_df









    Out[6]:






  
    
      
      power,active
      power,apparent
      energy,active
      energy,apparent
      voltage
    
  
  
    
      0 
        1
        2
        3
        4
        5
    
    
      10
       11
       12
       13
       14
       15
    
    
      20
       21
       22
       23
       24
       25

Save to HDF5



In [7]:

    
store = pd.HDFStore('column_test.h5', 'w')
store.put('string_df', string_df)
store.put('multiindex_df', multiindex_df)
store.close()

Read from HDF5



In [8]:

    
store = pd.HDFStore('column_test.h5', 'r')
store.get('multiindex_df')









    Out[8]:






  
    
      physical quantity
      power
      energy
      voltage
    
    
      type
      active
      apparent
      active
      apparent
      
    
  
  
    
      0 
        1
        2
        3
        4
        5
    
    
      10
       11
       12
       13
       14
       15
    
    
      20
       21
       22
       23
       24
       25



In [9]:

    
store.get('string_df')









    Out[9]:






  
    
      
      power,active
      power,apparent
      energy,active
      energy,apparent
      voltage
    
  
  
    
      0 
        1
        2
        3
        4
        5
    
    
      10
       11
       12
       13
       14
       15
    
    
      20
       21
       22
       23
       24
       25



In [10]:

    
store.close()

Dump to CSV



In [11]:

    
def print_csv(df):
    output_csv = StringIO()
    df.to_csv(output_csv)df.columns.set_names(LEVEL_NAMES, inplace=True)
    print(output_csv.getvalue())
    
print_csv(multiindex_df)









    



physical quantity,power,power,energy,energy,voltage
type,active,apparent,active,apparent,
,,,,,
0,1,2,3,4,5
10,11,12,13,14,15
20,21,22,23,24,25



In [12]:

    
print_csv(string_df)









    



,"power,active","power,apparent","energy,active","energy,apparent",voltage
0,1,2,3,4,5
10,11,12,13,14,15
20,21,22,23,24,25

Select data

Select single column



In [13]:

    
# MultiIndex
multiindex_df['power', 'active']









    Out[13]:





0      1
10    11
20    21
Name: (power, active), dtype: int64



In [14]:

    
# String
string_df['power,active']









    Out[14]:





0      1
10    11
20    21
Name: power,active, dtype: int64

select all power columns



In [15]:

    
multiindex_df['power'] # elegant way to get both 'active' and 'apparent' power data



In [16]:

    
# or like this:
multiindex_df['power']['active']









    Out[16]:





0      1
10    11
20    21
Name: active, dtype: int64



In [17]:

    
# String:
string_df[[col for col in string_df.columns if 'power' in col]]









    Out[17]:






  
    
      
      power,active
      power,apparent
    
  
  
    
      0 
        1
        2
    
    
      10
       11
       12
    
    
      20
       21
       22

list of all available 'power' AC types:



In [18]:

    
multiindex_df['power'].columns









    Out[18]:





Index([u'active', u'apparent'], dtype='object')



In [19]:

    
[col.split(',')[1] for col in string_df.columns if 'power' in col]









    Out[19]:





['active', 'apparent']

get all 'active' physical quantities



In [20]:

    
multiindex_df.xs('active', level='type', axis=1)









    Out[20]:






  
    
      physical quantity
      power
      energy
    
  
  
    
      0 
        1
        3
    
    
      10
       11
       13
    
    
      20
       21
       23



In [21]:

    
# String:
string_df[[col for col in string_df.columns if 'active' in col]]









    Out[21]:






  
    
      
      power,active
      energy,active
    
  
  
    
      0 
        1
        3
    
    
      10
       11
       13
    
    
      20
       21
       23

Iterate



In [26]:

    
for (physical_quantity, ac_type), series in multiindex_df.iteritems():
    print(physical_quantity, ac_type)









    



power active
power apparent
energy active
energy apparent
voltage



In [ ]:

physical quantity	power		energy		voltage
type	active	apparent	active	apparent
0	0.698756	1.366798	2.320818	-1.592383	0.155624
1	-0.817807	2.438316	1.541125	-0.393314	0.693330

	power,active	power,apparent	energy,active	energy,apparent	voltage
0	-0.979919	-0.431717	-0.073521	0.731000	0.109189
1	0.494156	0.410478	0.820294	-0.812069	0.256516