CODEWITHSUNDEEP
pythonpandas
ScalaBoy
ScalaBoy

Reputation: 3392

How to rename columns using the dictionary?

I have several pandas DataFrame's and I want to align their column names, so that all of them have the same names of particular columns (not all columns).

In my real data sets I have many columns, but below I provide a simplified example of 3 DataFrames. All of them have the same content, but it is done to simplify the example.

df1 =

col1   col2    col3
111    123     abc
122    331     zzz

df2 =

colA   colB    col3
111    123     abc
122    331     zzz

df3 =

col_1   col_2    col3
111     123      abc
122     331      zzz

Then I have the following dictionary that specifies similar columns (in reality the dictionary is bigger):

col_names = {
  "col1": ["colA", "col_1"],
  "col2": ["colB", "col_2"]
}

It means that the columns colA and col_1 should be renamed as col1, and the columns colB and col_2 should be renamed to col2.

I know how to rename columns one by one in pandas DataFrame:

df.rename(columns={"colA": "col1"}, inplace=True)

However, I am confused how to use the dictionary to rename columns flexibly?

Upvotes: 2

Views: 4088

Answers (2)

mujjiga
mujjiga

Reputation: 16876

df1.rename(columns={col_names[key][0]: key for key in col_names}, inplace=True)
df2.rename(columns={col_names[key][1]: key for key in col_names}, inplace=True)

If the order of values in the dictionary is random and also if you are not sure about columns in data frames then you can use.

df1 = pd.DataFrame({'col1': [1]*3, 'col2': [2]*3, 'col3': [3]*3})
df2 = pd.DataFrame({'colA': [11]*3, 'colB': [22]*3, 'col3': [33]*3})
df3 = pd.DataFrame({'col_1': ['a']*3, 'col_2': ['b']*3, 'col3': ['c']*3})
col_names = {
  "col1": ["colA", "col_1"],
  "col2": ["colB", "col_2"]
}
cols = {}
for key,value in col_names.items():
    for v in value:
        cols[v] =  key
for df in [df1,df2,df3]:
   df.rename(columns=cols, inplace=True)

Upvotes: 4

Rajat Jain
Rajat Jain

Reputation: 2032

Try:

df.columns = pd.Series(df.columns.to_list()).replace({'colA':'col1'}).to_list()

Upvotes: 0

Related Questions