pandas數(shù)據(jù)分組groupby()和統(tǒng)計(jì)函數(shù)agg()的使用

# 一個(gè)Series其實(shí)就是一條數(shù)據(jù)，Series方法的第一個(gè)參數(shù)是data，第二個(gè)參數(shù)是index(索引)，如果沒(méi)有傳值會(huì)使用默認(rèn)值（0-N）
# index參數(shù)是我們自定義的索引值，注意：參數(shù)值的個(gè)數(shù)一定要相同。
# 在創(chuàng)建Series時(shí)數(shù)據(jù)并不一定要是列表，也可以將一個(gè)字典傳進(jìn)去。
from pandas import Series, DataFrame

# 使用字典創(chuàng)建
index_list = ['001', '002', '003', '004', '005', '006', '007', '008', '009', '010']
name_list = ['李白', '王昭君', '諸葛亮', '狄仁杰', '孫尚香', '妲己', '周瑜', '張飛', '王昭君', '大喬']
age_list = [25, 28, 27, 25, 30, 29, 25, 32, 28, 26]
gender_list = ['F', 'M', 'F', 'F', 'M', 'M', 'F', 'F', 'M', 'M']
salary_list = ['10k', '12.5k', '20k', '14k', '12k', '17k', '18k', '21k', '22k', '21.5k']
marital_list = ['NO', 'NO', 'YES', 'YES', 'NO', 'NO', 'NO', 'YES', 'NO', 'YES']
dic = {
 '姓名': Series(data=name_list, index=index_list),
 '年齡': Series(data=age_list, index=index_list),
 '薪資': Series(data=salary_list, index=index_list),
 '性別': Series(data=gender_list, index=index_list),
 '婚姻狀況': Series(data=marital_list, index=index_list)
}
df = DataFrame(dic)

# 寫(xiě)入csv，path_or_buf為寫(xiě)入文本文件
df.to_csv(path_or_buf='./People.csv', encoding='utf_8_sig')
print('end')

import pandas as pd
df = pd.read_csv('./People.csv')
df.head()

# 根據(jù) '性別列' 進(jìn)行分組, 得到的是一個(gè)分組后的對(duì)象
groups = df.groupby('性別')
print(groups)

pandas.core.groupby.generic.DataFrameGroupBy object at 0x000002953DAEBC88>

print(groups.size())

性別
F 5
M 5
dtype: int64

group = df['年齡'].groupby(df['性別'])
# 查看分組
print(group.groups)
# 根據(jù)分組后的名字選擇分組
print(group.get_group('F'))

{'F': Int64Index([0, 2, 3, 6, 7], dtype='int64'), 'M': Int64Index([1, 4, 5, 8, 9], dtype='int64')}
0 25
2 27
3 25
6 25
7 32
Name: 年齡, dtype: int64

import pandas as pd
df = pd.read_csv('./People.csv')
groups = df.groupby('性別')
# print(groups)
for group_name,group_df in groups:
 print('分組的名稱(chēng):', group_name, '分組的數(shù)據(jù)', group_df.shape)
 print('-'*10)

import pandas as pd
df = pd.read_csv('./People.csv')
groups = df.groupby('性別')
for group_name,group_df in groups:
 f_mean = group_df['年齡'].mean()
 f_max = group_df['年齡'].max()
 f_min = group_df['年齡'].min()
 print('{}組的最大年齡是{}，最小年齡是{}，平均年齡是{}'.format(group_name,f_max,f_min,f_mean))

df = pd.read_csv('./People.csv')
group=df.groupby(['性別', '婚姻狀況'])
df1 = group.size()
print(df1)

性別 婚姻狀況
F  NO   2
  YES   3
M  NO   4
  YES   1
dtype: int64

size = df1['F'][ 'NO']
print(size)

2

import pandas as pd
df = pd.read_csv('./People.csv')
groups = df.groupby('性別')
for group_name,group_df in groups:
  f_se = group_df['年齡'].agg(['max','min','mean'])
  print('{}組的最大年齡是{}，最小年齡是{}，平均年齡是{}'.format(group_name,f_se[0],f_se[1],f_se[2]))

# 自定義的統(tǒng)計(jì)函數(shù)
def my_peak_range(df):
  """
  返回最大值與最小之間的范圍
  """
  return df.max() - df.min()

# 使用
for group_name,group_df in groups:
  f_se = group_df['年齡'].agg(['max','min','mean',my_peak_range])
  print(f_se[0],f_se[1],f_se[3])

32.0 25.0 7.0
30.0 26.0 4.0

import pandas as pd
df = pd.read_csv('./People.csv')
groups = df.groupby('性別')
gende=groups.sum().sort_values(by='年齡',ascending=False).index.to_list()[0]
"""
這行代碼, 先按性別進(jìn)行分組, 然后吧每組中的數(shù)據(jù)求和得到總的年齡, 在按照年齡排序
再取出index,最后使用to_list()轉(zhuǎn)換為列表,取出第一個(gè)數(shù)據(jù)
"""

print(gende)