Python 3.x - Create dataframe with column names from another dataframe

Question

Dataframe df1 contains a field 'Column headers' which has the column names. I want to create another dataframe df2 which only contains column headers from 'Column headers' column of df1.

print(df1['Column header'])
>>
0                                              % Female
1                                  % Below poverty line
2                                    % Rural population
3                      Decadal Population Growth (in %)
4     Availability of Drinking Water Source Within P...
5                                 Concrete Roofs (in %)
6                        Houses With Electricity (in %)
7                        Houses With Televisions (in %)
8                           With Computer/Laptop (in %)
9        Houses With Phones (Telephone + Mobile) (in %)
10                        Houses With 2 wheelers (in %)
11                              Houses With cars (in %)
12              Households With Banking Services (in %)
13                                 Literacy Rate (in %)
14                         Literacy Rate (Rural) (in %)
15                         Literacy Rate (Urban) (in %)
16                  Decadal Difference In Literacy Rate
17                 Student: Teacher Ratio - All Schools
18                     Student: Teacher Ratio - Primary
19               Student: Teacher Ratio - Upper Primary
20     Under-five Mortality Rate (Per 1000 live Births)
21           No of Dispensaries per 1,00,000 population
22                No of Doctors per 1,00,000 population
23    Total patients registered for tuberculosis tre...
24                   Sex Ratio (Females Per 1000 Males)
25                                        Agri GSDP (%)
26                                    Industry GSDP (%)
27                                     Service GSDP (%)
28                        Unemployment Rate   (2011-12)
29                   Rural Unemployment Rate  (2011-12)
30                   Urban Unemployment Rate  (2011-12)
31                Per Capita Public Expenditure (in Rs)
32               Per Capita Private Expenditure (in Rs)
33                          Infant Mortality Rate (IMR)
34                             Maternal Mortality Rate 
35          Coverage Of National Highways (Total in km)
36             Coverage Of State Highways (Total in km)
37                Coverage Of Rural Roads (Total in km)
38                Coverage Of Urban Roads (Total in km)
39                       Railway Coverage (Total in km)
40    Tele-Density [Total Connections /  Total Popul...
Name: Column headers, dtype: object

I want to create dataframe df2 which contains 40 columns as mentioned above. The rows in this dataframe will be populated by a different function. I tried to create df2 as follows -

df2 = pd.DataFrame()  #Creating an empty dataframe 
df2.columns = df1['Column header']
>>
ValueError: Length mismatch: Expected axis has 0 elements, new values have 41 elements

Is it possible to create a blank dataframe in Pandas and specify the column names afterwards?

Solution

try this:

df2 = pd.DataFrame(columns=df1['Column header'])

but you shouldn't create empty DFs, because it's very slow to fill them up row by row. So you should collect your data first and then create your DF using precollected data.

Comments

[Rohan Bapat]

Exactly what i was looking for. Thanks!