[kaggle] Learn Tutorial_Pandas (정리)_2

< 5 > Data Types and Missing Values

this is about) how to investigate data types within a DataFrame or Series

also going to learn about)how to find and replace entries.

 

Dtypes

What is Dtype? The data type for a column in a DataFrame or a Series

1. dtype, dtypes, astype()

- dtypes: returns the dtype of every column in the DataFrame

- keep in mind) columns consisting entirely of strings do not get their own type; they are given the object type

reviews.price.dtype			# dtype('float64')
reviews.dtypes     #shows each coloumns' dtype

'''
country                   object
description               object
designation               object
points                     int64
price                    float64
province                  object
region_1                  object
region_2                  object
taster_name               object
taster_twitter_handle     object
title                     object
variety                   object
winery                    object
dtype: object
'''

 

- astype()

: to convert a column of one type into another

reviews.points.astype('float64')

 

- even index of DataFrame or Series has its own dtype

reviews.index.dtype			# dtype('int64')

 

 

 

Missing Data

NaN: "Not a Number", 'float64' dtype

1. pd.isnull()

: to select NaN values

reviews[pd.isnull(reviews.country)]     #find if 'country' value is NaN

 

2. pd.notnull()

: to select not NaN values

reviews[pd.notnull(reviews.country)]

 

3. fillna()

: Replacing missing values

reviews.region_2.fillna("Unknown")

 

4. replace("A", "B")   // replace "A" to "B"

: Used at non-null value that we would like to replace

reviews.taster_twitter_handle.replace("@kerinokeefe", "@kerino")

 

 

< 6 > Renaming and Combining

Renaming

1.

(1) rename(columns={'A': 'B'})               # column 이름 A를 B로 변경

: lets you change index names and/or column names

rename index or column values by specifying an index or column keyword parameter, respectively

reviews.rename(columns={'points': 'score'})

 

(2) rename(index={0 : 'firstEntry', 1 : 'secondEntry'})       #row index 숫자를 문자열로 변경

reviews.rename(index={0:'firstEntry', 1:'secondEntry'})

 

(3) rename_axis("name_you_want", axis='rows or columns')

: Both the row index and the column index can have their own name attribute

reviews.rename_axis("fields", axis='columns').rename_axis("wines", axis='rows')

 

Combining

1. concat()

: The simplest combining method

This is useful when we have data in different DataFrame or Series objects but have the same fields (columns).

canadian_youtube = pd.read_csv("/content/drive/MyDrive/Kaggle/project0/CAvideos.csv")
british_youtube = pd.read_csv("/content/drive/MyDrive/Kaggle/project0/GBvideos.csv")

pd.concat([canadian_youtube, british_youtube])

 

2. join()

: The middlemost combiner in terms of complexity

combine different DataFrame objects which have an index in common.

 

parameters

lsuffix : 중복된 column이 있을 경우 left DataFrame의 column명에 붙일 suffix

rsuffix : 중복된 column이 있을 경우 right DataFrame의 column명에 붙일 suffix

left = canadian_youtube.set_index(['title', 'trending_date'])
right = british_youtube.set_index(['title', 'trending_date'])

left.join(right, lsuffix='_CAN', rsuffix='_UK')