数据分析-day03-pandas-dataFrame的统计函数

# -*- coding: utf-8 -*-

# @File    : pandas_dataframe_total_demo.py
# @Date    :  2020-01-04 11:52
# @Author  : admin
'''
统计函数
'''
import pandas as pd
import numpy as np

file_path = "../../data/IMDB-Movie-Data.csv"
df = pd.read_csv(file_path)
#dataframe获取某行某列的数据
print(df.loc[1,["Actors"]])
# print(df.info())
print(df.head(1))
#print(df.iloc[1])
print(df.columns)
#获取平均评分
print("rating的平均分:",df["Rating"].mean())

#导演的人数
# print(len(set(df["Director"].tolist())))
print(len(df["Director"].unique()))
print(len(df["Director"].unique()))

#dataframe获取某行某列的数据
print(df.loc[1,["Actors"]])

#截取字符串，并统计演员的人数
temp_list=df["Actors"].str.split(",").tolist()
actor_list=[n for m in temp_list for n in m];
actor_num=len(set(actor_list))
print("actor 的数目:",actor_num)

结果：

Actors    Noomi Rapace, Logan Marshall-Green, Michael Fa...
Name: 1, dtype: object
   Rank                    Title  ... Revenue (Millions) Metascore
0     1  Guardians of the Galaxy  ...             333.13      76.0

[1 rows x 12 columns]
Index(['Rank', 'Title', 'Genre', 'Description', 'Director', 'Actors', 'Year',
       'Runtime (Minutes)', 'Rating', 'Votes', 'Revenue (Millions)',
       'Metascore'],
      dtype='object')
rating的平均分: 6.723199999999999
644
644
Actors    Noomi Rapace, Logan Marshall-Green, Michael Fa...
Name: 1, dtype: object
actor 的数目: 2394