Pandas: get value of an index in multi-index-CodePudding

I have a pandas dataframe :

data = {'id': [1,1,2,3],
        'name': ["Noemie","Elena", "Lize", "Patrick"],
        'value': ["pink", "blue", "red", "yellow" ]}

So I gather all the rows with the id "1" thanks to df.set_index(["id","name"])

My columns id and name become indexes, I wonder if is it possible to get for example the value "Noemie"

Knowing that, before multi-indexing I can have it with

df.loc[0].at["name"]

CodePudding user response：

You can use df.index.get_level_values to get the desired index column you want and do comparison.

df[df.index.get_level_values('name') == 'Noemie']
 
          value
id name        
1  Noemie  pink

Or if you just want the index value

df.index[df.index.get_level_values('name') == 'Noemie'][0][1]
Out[30]: 'Noemie'

CodePudding user response：

If need select first row by DataFrame.iloc then get second value of indices by name[1] - python count from 0:

data = {'id': [1,1,2,3],
        'name': ["Noemie","Elena", "Lize", "Patrick"],
        'value': ["pink", "blue", "red", "yellow" ]}

df = pd.DataFrame(data).set_index(["id","name"])
print (df.iloc[0])
value    pink
Name: (1, Noemie), dtype: object

print (df.iloc[0].name)
(1, 'Noemie')

out = df.iloc[0].name[1]
print (out)
Noemie

For select second row:

print (df.iloc[1])
value    blue
Name: (1, Elena), dtype: object

print (df.iloc[1].name)
(1, 'Elena')

out = df.iloc[1].name[1]
print (out)
Elena

EDIT: If need all values of index if id=1 use:

print (df.loc[1])
       value
name        
Noemie  pink
Elena   blue

s = df.loc[1].index.tolist()
print (s)
['Noemie', 'Elena']