[pandas] add new column / delete column

1. Column

import pandas as pd
df = pd.DataFrame({'a': [1,1,3,4,5], 'b': [2,3,2,3,4], 'c': [3,4,7,6,4]})

# 1. to simply add new column
df['d'] = [1,3,6,4,8]

# 2. add one number in that column (will be filled out with that number)
df['e'] = 1

# +) calculated result can be also created as a new column.
# check datatype first
print(df.dtypes)
df['f'] = df['a'] + df['b'] - df['c']

# delete a column
df.drop(['d', 'e', 'f'], axis=1, inplace=True)

when you delete a column, be aware that you need to write 'axis=1' to notify program that the user wants to delete the column.

 

2. Row

import pandas as pd
df = pd.DataFrame({'a': [1,1,3,4,5], 'b': [2,3,2,3,4], 'c': [3,4,7,6,4]})

# simply add a new row - 'ignore_index=True' should be written!!!
df = df.append({'a':6, 'b':7, 'c':8}, ignore_index=True)

# add a new row by using loc (it will be added after the last row of the original dataset)
df.loc[6] = [7,8,9]
# however, if the number n in loc[n] is between the existing dataframe,
# then the new row data will be replaced that existing data
df.loc[1] = [7,8,9]

there are several ways to delete/drop the row of the data 

# delete a certain row - similar to list index info. df.drop(1) := drop/delete the second row
df = df.drop(1)
print(df)

# delete several rows in a time
df = df.drop([0,1])
print(df)

# delete a range of rows in a time
# limitation: only can delete from first row to nth row
df = df.drop([i for i in range(4)])

# more flexible way
df = df.drop(df[df['a'] < 4].index)
# more than one conditions
df = df.drop(df[(df['a'] < 3) & (df['c'] == 4)].index)