Scribd
Upload
4 views

pandas-1

The document provides an overview of using the pandas library in Python for data manipulation, including creating Series and DataFrames, merging, concatenating, and joining data. It demonstrates how to create Series from lists and dictionaries, perform various types of merges (inner, left, right, outer), and read CSV files into DataFrames. Additionally, it emphasizes the importance of data types in DataFrames and how to handle them effectively.

Uploaded by

praveen838307
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
4 views

pandas-1

The document provides an overview of using the pandas library in Python for data manipulation, including creating Series and DataFrames, merging, concatenating, and joining data. It demonstrates how to create Series from lists and dictionaries, perform various types of merges (inner, left, right, outer), and read CSV files into DataFrames. Additionally, it emphasizes the importance of data types in DataFrames and how to handle them effectively.

Uploaded by

praveen838307
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

pandas-1

April 21, 2025

[1]: import pandas as pd

[ ]: # it is created using a list / tuple

[2]: sr1 = pd.Series([10,20,30,40,50])


sr1

[2]: 0 10
1 20
2 30
3 40
4 50
dtype: int64

[3]: type(sr1)

[3]: pandas.core.series.Series

[4]: sr1[3]

[4]: 40

[ ]: # pandas.core.series.Series

[ ]: # pd.Series --> Series is a function which is available in pandas module

[6]: list1 = [10,20,30,40,50]


list1[3]

[6]: 40

[5]: sr2 = pd.Series([10,20,30,40,50],index = ['a','b','c','d','e'], dtype = 'float')


sr2

[5]: a 10.0
b 20.0
c 30.0

1
d 40.0
e 50.0
dtype: float64

[7]: sr2['d']

[7]: 40.0

[ ]:

DataFrame
[10]:

[10]: 0 1
1 2
2 3
3 4
dtype: int64

[8]: # creating a dataframe using a dictionary is the most easiest way


# data frame --> table --> rows and columns
# columns can be anything : list , tuple,set , array(1D), series
dict1 = {
'col1': [10,20,30,40,50],
'col2': pd.Series(['a','b','c','d','e'])
}
df1 = pd.DataFrame(dict1)
df1
# each column should have single data type
# col1 can have either int data --> it cant process [1,'a',2.0]
# each column can have its own data type (single data type)

[8]: col1 col2


0 10 a
1 20 b
2 30 c
3 40 d
4 50 e

[12]: df1.dtypes
# int --> int64/32
# float --> float64/float32
# string --> objects

[12]: col1 int64


col2 object

2
dtype: object

[ ]:

Merge
[13]: df1 = pd.DataFrame({'id': [1,2,3,4,5], 'ml_marks': [94,95,98,92,90]})
df1

[13]: id ml_marks
0 1 94
1 2 95
2 3 98
3 4 92
4 5 90

[14]: df2 = pd.DataFrame({'id': [3,4,5,6,7], 'stat_marks': [98,84,88,96,94]})


df2

[14]: id stat_marks
0 3 98
1 4 84
2 5 88
3 6 96
4 7 94

[15]: df1.merge(df2, on = 'id',how = 'inner')

[15]: id ml_marks stat_marks


0 3 98 98
1 4 92 84
2 5 90 88

[16]: df1.merge(df2, on = 'id',how = 'left')

[16]: id ml_marks stat_marks


0 1 94 NaN
1 2 95 NaN
2 3 98 98.0
3 4 92 84.0
4 5 90 88.0

[17]: df1.merge(df2, on = 'id',how = 'right')

[17]: id ml_marks stat_marks


0 3 98.0 98
1 4 92.0 84

3
2 5 90.0 88
3 6 NaN 96
4 7 NaN 94

[18]: df1.merge(df2, on = 'id',how = 'outer') # it will return every record

[18]: id ml_marks stat_marks


0 1 94.0 NaN
1 2 95.0 NaN
2 3 98.0 98.0
3 4 92.0 84.0
4 5 90.0 88.0
5 6 NaN 96.0
6 7 NaN 94.0

[ ]: # df1.merge(df2,how = 'cross')
# df1.merge(df1, on = 'id')

[ ]:

Concate
[ ]: # combine two data frame at row level ( increaase no.of rows by using second␣
↪data frame)

[19]: df1 = pd.DataFrame({'id': [1,2,3,4,5], 'ml_marks': [94,95,98,92,90]})


df2 = pd.DataFrame({'id': [6,7,8,9,10], 'ml_marks': [94,95,98,92,90]})

[20]: df2

[20]: id ml_marks
0 6 94
1 7 95
2 8 98
3 9 92
4 10 90

[21]: pd.concat([df1,df2])

[21]: id ml_marks
0 1 94
1 2 95
2 3 98
3 4 92
4 5 90
0 6 94
1 7 95

4
2 8 98
3 9 92
4 10 90

[ ]: # as in reality data is not stored in sequential manner joining on index doesnt␣


↪make any sense

[22]: df1 = pd.DataFrame({'id': [1,2,3,4,5], 'ml_marks': [94,95,98,92,90]})


df2 = pd.DataFrame({'id': [3,4,5,6,7], 'stat_marks': [98,84,88,96,94]})

[ ]: # Join --> it performs the same operation of combining data frames like Merge␣
↪but on index values

# join is performed on index values


# suffix parameter is needed to differentiate common column names

[23]: # snake_case : divide larger words by using _ --> new_df


# NewDf
# newDf
df1.join(df2, lsuffix = '_df1',rsuffix='_df2')

[23]: id_df1 ml_marks id_df2 stat_marks


0 1 94 3 98
1 2 95 4 84
2 3 98 5 88
3 4 92 6 96
4 5 90 7 94

[ ]: # download Auto.csv from drive folder

[24]: pwd # folder where your current code is available


# pandas-1 is available in 2802

[24]: 'C:\\Users\\admin\\2802'

[ ]: # copy paste Auto.csv into the folder 'C:\\Users\\admin\\2802'


# ensure the code and data set should be in same folder

[ ]: # pd.read_excel('filename.xlsx')

[25]: df1 = pd.read_csv('Auto.csv')


df1.head()

[25]: mpg cylinders displacement Horse Power weight acceleration year \


0 18.0 8.0 307.0 130 3504 12.0 70
1 15.0 8.0 350.0 165 3693 11.5 70
2 NaN 8.0 318.0 150 3436 11.0 70
3 NaN 8.0 NaN 150 3433 12.0 70

5
4 NaN 8.0 NaN 140 3449 10.5 70

origin name
0 1 chevrolet chevelle malibu
1 1 buick skylark 320
2 1 plymouth satellite
3 1 amc rebel sst
4 1 ford torino

You might also like