Slicing Dataframe using Loc and iLoc

Contents

3.5. Slicing Dataframe using Loc and iLoc#

import pandas as pd
import numpy as np

df=pd.DataFrame({
'Name':['Sahil','Sonia','Sourav','Vishal'],
'Age':[10,20,30,40],
'Gender':['M','F','M','M'],
'City':['J','K','L','P'],
'Work':[True,False,False,True]
}
)
df

	Name	Age	Gender	City	Work
0	Sahil	10	M	J	True
1	Sonia	20	F	K	False
2	Sourav	30	M	L	False
3	Vishal	40	M	P	True

3.5.1. Accessing columns by Names#

3.5.1.1. Accessing single column#

df['Name'] # Series

0     Sahil
1     Sonia
2    Sourav
3    Vishal
Name: Name, dtype: object

# Getting the value of column
df['Name'][2]

'Sourav'

df[['Name']] # Dataframe

	Name
0	Sahil
1	Sonia
2	Sourav
3	Vishal

df.Name  # Another way

0     Sahil
1     Sonia
2    Sourav
3    Vishal
Name: Name, dtype: object

3.5.1.2. Accessing multiple columns#

df[['Name','Age','Gender']]

	Name	Age	Gender
0	Sahil	10	M
1	Sonia	20	F
2	Sourav	30	M
3	Vishal	40	M

# Another way
list_cols=['Name','Gender']
df[list_cols]

	Name	Gender
0	Sahil	M
1	Sonia	F
2	Sourav	M
3	Vishal	M

3.5.1.3. Using loc#

df.loc[rows,columns] # Note:square brackets are used with loc
- all rows or filtered rows
- example, df.loc[df[‘age’]>50,’name’] === df[df[‘age’][‘name’]
- df.loc[0,:]
- df.loc[[1,2,3],:]
- df.loc[[1,2,3]:[‘Name,’Age’,’Gender’]

df.loc[:,['Name','Gender']]

	Name	Gender
0	Sahil	M
1	Sonia	F
2	Sourav	M
3	Vishal	M

3.5.2. Accessing range of columns#

3.5.2.1. By index#

df.iloc[:,[1,4]]

	Age	Work
0	10	True
1	20	False
2	30	False
3	40	True

3.5.2.2. By Name#

df.loc[:,['Name','Gender']]

	Name	Gender
0	Sahil	M
1	Sonia	F
2	Sourav	M
3	Vishal	M

df.iloc[:,1:4]

	Age	Gender	City
0	10	M	J
1	20	F	K
2	30	M	L
3	40	M	P

df.loc[:,'Name':'Gender']

	Name	Age	Gender
0	Sahil	10	M
1	Sonia	20	F
2	Sourav	30	M
3	Vishal	40	M

3.5.2.3. Give all columns of 0th row#

df.loc[0,:]

Name      Sahil
Age          10
Gender        M
City          J
Work       True
Name: 0, dtype: object

3.5.2.4. Give all columns of first 3 rows#

df.loc[[0,1,2],:]

	Name	Age	Gender	City	Work
0	Sahil	10	M	J	True
1	Sonia	20	F	K	False
2	Sourav	30	M	L	False

3.5.3. Notes#

loc is used to access through names
iloc is used to access through indexes
, in used to select given columns
: is used to select range of columns
if you want particular columns during fetching dataframe,use columns=list of columns you want to select

3.5.4. Summary#

# Get first row 
df.loc[0]

Name      Sahil
Age          10
Gender        M
City          J
Work       True
Name: 0, dtype: object

# Get first two rows 
df.loc[0:1] 

	Name	Age	Gender	City	Work
0	Sahil	10	M	J	True
1	Sonia	20	F	K	False

# Get first two rows's specific column
df.loc[0:1].Name

0    Sahil
1    Sonia
Name: Name, dtype: object

# Get even numbered rows
df.loc[df.index%2==0]

	Name	Age	Gender	City	Work
0	Sahil	10	M	J	True
2	Sourav	30	M	L	False

# Get odd numbered rows
df.loc[df.index%2!=0]

	Name	Age	Gender	City	Work
1	Sonia	20	F	K	False
3	Vishal	40	M	P	True

# Add steps in rows
# Select every nth row
# df[df.index % n == 0]  # Selects every nth raw starting from 0
df[df.index % 2 == 0]

	Name	Age	Gender	City	Work
0	Sahil	10	M	J	True
2	Sourav	30	M	L	False

# Excludes every nth row
df[df.index % 3 != 0]

	Name	Age	Gender	City	Work
1	Sonia	20	F	K	False
2	Sourav	30	M	L	False