005.python科学计算库pandas(下)

怼烎@ 2021-10-06 15:18 460阅读 0赞

### 测试数据 [fandango\_score\_comparison.csv][fandango_score_comparison.csv] ###

--------------------

### series ###

import pandas as pd from pandas import Series fandango = pd.read_csv('fandango_score_comparison.csv') series_film = fandango['FILM'] print(series_film[0:5]) print("----------------------------------") series_rt = fandango['RottenTomatoes'] print(series_rt[0:5]) print("----------------------------------") film_names = series_film.values rt_scores = series_rt.values # 带轴标的一维ndarray(包括时间序列)。 series_custom = Series(rt_scores, index=film_names) print(series_custom[['Minions (2015)', 'Leviathan (2014)']]) print("----------------------------------") print(series_custom[4:6])

![pycharm64_2018-09-14_17-57-28.png][]

--------------------

### sort ###

import pandas as pd from pandas import Series fandango = pd.read_csv('fandango_score_comparison.csv') series_film = fandango['FILM'] series_rt = fandango['RottenTomatoes'] film_names = series_film.values rt_scores = series_rt.values # 带轴标的一维ndarray(包括时间序列)。 series_custom = Series(rt_scores, index=film_names) original_index = series_custom.index.tolist() # sorted 以升序返回一个包含迭代中所有项的新列表。 sorted_index = sorted(original_index) print(sorted_index) print("----------------------------------") # 按照已排序后的sorted_index来排序series_custom (sorted_index 和 series_custom.index 元素需保持一致) sorted_by_index = series_custom.reindex(sorted_index) print(sorted_by_index)

![pycharm64_2018-09-14_18-10-56.png][]

![pycharm64_2018-09-14_18-12-47.png][]

--------------------

### series算术运算 ###

import pandas as pd from pandas import Series import numpy as np fandango = pd.read_csv('fandango_score_comparison.csv') series_film = fandango['FILM'] series_rt = fandango['RottenTomatoes'] film_names = series_film.values rt_scores = series_rt.values # 带轴标的一维ndarray(包括时间序列)。 series_custom = Series(rt_scores, index=film_names) print(series_custom[0:3]) print("----------------------------------") # 将每个值相加 print(np.add(series_custom, series_custom)[0:3]) print("----------------------------------") # 对每个值应用正弦函数 print(np.sin(series_custom)[0:3]) print("----------------------------------") # 返回最高值(将返回单个值而不是一系列值) print(np.max(series_custom))

![pycharm64_2018-09-14_18-35-22.png][]

import pandas as pd from pandas import Series fandango = pd.read_csv('fandango_score_comparison.csv') series_film = fandango['FILM'] series_rt = fandango['RottenTomatoes'] film_names = series_film.values rt_scores = series_rt.values # 带轴标的一维ndarray(包括时间序列)。 series_custom = Series(rt_scores, index=film_names) # series_custom > 50 会为每个film返回一个带有布尔值的系列对象 series_greater_than_50 = series_custom[series_custom > 50] print(series_greater_than_50[0:5]) print("-----------------------------------") both_criteria = series_custom[(series_custom > 50) & (series_custom < 75)] print(both_criteria[0:5])

![pycharm64_2018-09-14_18-42-52.png][]

import pandas as pd from pandas import Series fandango = pd.read_csv('fandango_score_comparison.csv') rt_critics = Series(fandango['RottenTomatoes'].values, index=fandango['FILM']) print(rt_critics[0:3]) rt_users = Series(fandango['RottenTomatoes_User'].values, index=fandango['FILM']) print(rt_users[0:3]) rt_mean = (rt_critics + rt_users) / 2 print(rt_mean[0:3])

![pycharm64_2018-09-14_18-45-56.png][]

--------------------

### set\_index ###

import pandas as pd from pandas import Series fandango = pd.read_csv('fandango_score_comparison.csv') # set_index 使用一个或多个现有列设置DataFrame索引(行标签)。默认情况下，生成一个新对象。 # drop : boolean, default True 删除要用作新索引的列 fandango_films = fandango.set_index('FILM', drop=False) print(fandango_films[0:3])

![pycharm64_2018-09-14_18-52-06.png][]

*  当选择多个行时，返回一个DataFrame， 但当选择单个行时，返回的是一个Series对象

import pandas as pd fandango = pd.read_csv('fandango_score_comparison.csv') fandango_films = fandango.set_index('FILM', drop=False) # 使用括号表示法或loc[]进行切片 sub_films = fandango_films["Avengers: Age of Ultron (2015)":"Ant-Man (2015)"] print(sub_films) print(type(sub_films)) print("----------------------------------------------------") sub_films = fandango_films.loc["Avengers: Age of Ultron (2015)":"Ant-Man (2015)"] print(sub_films) print(type(sub_films)) print("----------------------------------------------------") # 查找特定的 movie film = fandango_films.loc['Kumiko, The Treasure Hunter (2015)'] print(type(film)) print("----------------------------------------------------") # 查找特定的 movie 列表 movies = ['Kumiko, The Treasure Hunter (2015)', 'Do You Believe? (2015)'] print(fandango_films.loc[movies]) print(type(fandango_films.loc[movies])) # 当选择多个行时，返回一个DataFrame， # 但当选择单个行时，返回的是一个Series对象

![pycharm64_2018-09-15_15-28-53.png][]

import pandas as pd import numpy as np fandango = pd.read_csv('fandango_score_comparison.csv') fandango_films = fandango.set_index('FILM', drop=False) # panda中的apply()方法允许我们指定Python逻辑 # apply()方法需要传入一个矢量化操作 # 可以应用于每个系列对象。 # 以Series的形式返回数据类型 types = fandango_films.dtypes # print(types) 返回所有的列名称 print(type(types)) print("----------------------------------------------------") # 过滤数据类型为floats，索引属性只返回列名 float_columns = types[types.values == 'float64'].index # 使用括号表示法过滤列，使其只是float列 float_df = fandango_films[float_columns] print(float_df[0:3]) print(type(float_df)) print("----------------------------------------------------") # “x”是表示列的系列对象 # numpy.std 计算沿指定轴的标准差。 deviations = float_df.apply(lambda x: np.std(x)) print(deviations[0:3]) print(type(deviations)) print("----------------------------------------------------") rt_mt_user = float_df[['RT_user_norm', 'Metacritic_user_nom']] deviations = rt_mt_user.apply(lambda x: np.std(x), axis=1) print(deviations[0:3])

![pycharm64_2018-09-15_15-45-26.png][]

--------------------

[fandango_score_comparison.csv]: http://upyun.qubianzhong.cn/file/python_pandas/fandango_score_comparison.csv
[pycharm64_2018-09-14_17-57-28.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-14_17-57-28.png
[pycharm64_2018-09-14_18-10-56.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-14_18-10-56.png
[pycharm64_2018-09-14_18-12-47.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-14_18-12-47.png
[pycharm64_2018-09-14_18-35-22.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-14_18-35-22.png
[pycharm64_2018-09-14_18-42-52.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-14_18-42-52.png
[pycharm64_2018-09-14_18-45-56.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-14_18-45-56.png
[pycharm64_2018-09-14_18-52-06.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-14_18-52-06.png
[pycharm64_2018-09-15_15-28-53.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-15_15-28-53.png
[pycharm64_2018-09-15_15-45-26.png]: http://upyun.qubianzhong.cn/picture/pycharm64_2018-09-15_15-45-26.png