第01章 Pandas基础
第02章 DataFrame基础运算
第03章 创建和持久化DataFrame
第04章 开始数据分析
第05章 探索性数据分析
第06章 选取数据子集
第07章 过滤行
第08章 索引对齐
6.1 选取Series数据
读取大学数据集，使用校名作为索行引：
>>> import pandas as pd
>>> import numpy as np
>>> college = pd.read_csv(
... "data/college.csv", index_col="INSTNM"
... )
>>> city = college["CITY"]
>>> city
INSTNM
Alabama A & M University Normal
University of Alabama at Birmingham Birmingham
Amridge University Montgomery
University of Alabama in Huntsville Huntsville
Alabama State University Montgomery
...
SAE Institute of Technology San Francisco Emeryville
Rasmussen College - Overland Park Overland...
National Personal Training Institute of Cleveland Highland...
Bay Area Medical Academy - San Jose Satellite Location San Jose
Excel Learning Center-San Antonio South San Antonio
Name: CITY, Length: 7535, dtype: object
从Series选取标量值：
>>> city["Alabama A & M University"]
'Normal'
使用.loc提取标量值：
>>> city.loc["Alabama A & M University"]
'Normal'
使用.iloc提取标量值：
>>> city.iloc[0]
'Normal'
提取出多个值，如果传入的是列表，返回的是Series：
>>> city[
... [
... "Alabama A & M University",
... "Alabama State University",
... ]
... ]
INSTNM
Alabama A & M University Normal
Alabama State University Montgomery
Name: CITY, dtype: object
使用.loc重复上面的步骤：
>>> city.loc[
... [
... "Alabama A & M University",
... "Alabama State University",
... ]
... ]
INSTNM
Alabama A & M University Normal
Alabama State University Montgomery
Name: CITY, dtype: object
使用.iloc重复上面的步骤：
>>> city.iloc[[0, 4]]
INSTNM
Alabama A & M University Normal
Alabama State University Montgomery
Name: CITY, dtype: object
使用切片提取多值：
>>> city[
... "Alabama A & M University":"Alabama State University"
... ]
INSTNM
Alabama A & M University Normal
University of Alabama at Birmingham Birmingham
Amridge University Montgomery
University of Alabama in Huntsville Huntsville
Alabama State University Montgomery
Name: CITY, dtype: object
使用位置切片提取多值：
>>> city[0:5]
INSTNM
Alabama A & M University Normal
University of Alabama at Birmingham Birmingham
Amridge University Montgomery
University of Alabama in Huntsville Huntsville
Alabama State University Montgomery
Name: CITY, dtype: object
使用布尔数组提取多值：
>>> alabama_mask = city.isin(["Birmingham", "Montgomery"])
>>> city[alabama_mask]
INSTNM
University of Alabama at Birmingham Birmingham
Amridge University Montgomery
Alabama State University Montgomery
Auburn University at Montgomery Montgomery
Birmingham Southern College Birmingham
...
Fortis Institute-Birmingham Birmingham
Hair Academy Montgomery
Brown Mackie College-Birmingham Birmingham
Nunation School of Cosmetology Birmingham
Troy University-Montgomery Campus Montgomery
Name: CITY, Length: 26, dtype: object
更多
使用.loc和.iloc在原始DataFrame操作也可以实现同样的目的：
>>> college.loc["Alabama A & M University", "CITY"]
'Normal'
>>> college.iloc[0, 0]
'Normal'
>>> college.loc[
... [
... "Alabama A & M University",
... "Alabama State University",
... ],
... "CITY",
... ]
INSTNM
Alabama A & M University Normal
Alabama State University Montgomery
Name: CITY, dtype: object
>>> college.iloc[[0, 4], 0]
INSTNM
Alabama A & M University Normal
Alabama State University Montgomery
Name: CITY, dtype: object
>>> college.loc[
... "Alabama A & M University":"Alabama State University",
... "CITY",
... ]
INSTNM
Alabama A & M University Normal
University of Alabama at Birmingham Birmingham
Amridge University Montgomery
University of Alabama in Huntsville Huntsville
Alabama State University Montgomery
Name: CITY, dtype: object
>>> college.iloc[0:5, 0]
INSTNM
Alabama A & M University Normal
University of Alabama at Birmingham Birmingham
Amridge University Montgomery
University of Alabama in Huntsville Huntsville
Alabama State University Montgomery
Name: CITY, dtype: object
使用.loc切片时要注意，索引如果越界，会返回空值：
>>> city.loc[
... "Reid State Technical College":"Alabama State University"
... ]
Series([], Name: CITY, dtype: object)
6.2 选取DataFrame行
这一节和上节有点像，还是先读取数据：
>>> college = pd.read_csv(
... "data/college.csv", index_col="INSTNM"
... )
>>> college.sample(5, random_state=42)
CITY STABBR ... MD_EARN_WNE_P10 GRAD_DEBT_MDN_SUPP
INSTNM ...
Career Po... San Antonio TX ... 20700 14977
Ner Israe... Baltimore MD ... PrivacyS... PrivacyS...
Reflectio... Decatur IL ... NaN PrivacyS...
Capital A... Baton Rouge LA ... 26400 PrivacyS...
West Virg... Montgomery WV ... 43400 23969
<BLANKLINE>
[5 rows x 26 columns]
使用.iloc提取一整行：
>>> college.iloc[60]
CITY Anchorage
STABBR AK
HBCU 0
MENONLY 0
WOMENONLY 0
...
PCTPELL 0.2385
PCTFLOAN 0.2647
UG25ABV 0.4386
MD_EARN_WNE_P10 42500
GRAD_DEBT_MDN_SUPP 19449.5
Name: University of Alaska Anchorage, Length: 26, dtype: object
使用.loc实现上一步：
>>> college.loc["University of Alaska Anchorage"]
CITY Anchorage
STABBR AK
HBCU 0
MENONLY 0
WOMENONLY 0
...
PCTPELL 0.2385
PCTFLOAN 0.2647
UG25ABV 0.4386
MD_EARN_WNE_P10 42500
GRAD_DEBT_MDN_SUPP 19449.5
Name: University of Alaska Anchorage, Length: 26, dtype: object
使用.iloc提取一组不连续的行：
>>> college.iloc[[60, 99, 3]]
CITY STABBR ... MD_EARN_WNE_P10 GRAD_DEBT_MDN_SUPP
INSTNM ...
Universit... Anchorage AK ... 42500 19449.5
Internati... Tempe AZ ... 22200 10556
Universit... Huntsville AL ... 45500 24097
<BLANKLINE>
[3 rows x 26 columns]
使用.loc提取一组不连续的行：
>>> labels = [
... "University of Alaska Anchorage",
... "International Academy of Hair Design",
... "University of Alabama in Huntsville",
... ]
>>> college.loc[labels]
CITY STABBR ... MD_EARN_WNE_P10 GRAD_DEBT_MDN_SUPP
INSTNM ...
Universit... Anchorage AK ... 42500 19449.5
Internati... Tempe AZ ... 22200 10556
Universit... Huntsville AL ... 45500 24097
<BLANKLINE>
[3 rows x 26 columns]
使用.iloc提取一组连续的行：
>>> college.iloc[99:102]
CITY STABBR ... MD_EARN_WNE_P10 GRAD_DEBT_MDN_SUPP
INSTNM ...
Internati... Tempe AZ ... 22200 10556
GateWay C... Phoenix AZ ... 29800 7283
Mesa Comm... Mesa AZ ... 35200 8000
<BLANKLINE>
[3 rows x 26 columns]
.loc的切片是包含起始和结束的索引的：
>>> start = "International Academy of Hair Design"
>>> stop = "Mesa Community College"
>>> college.loc[start:stop]
CITY STABBR ... MD_EARN_WNE_P10 GRAD_DEBT_MDN_SUPP
INSTNM ...
Internati... Tempe AZ ... 22200 10556
GateWay C... Phoenix AZ ... 29800 7283
Mesa Comm... Mesa AZ ... 35200 8000
<BLANKLINE>
[3 rows x 26 columns]
更多
将行索引的序号转变为字符串：
>>> college.iloc[[60, 99, 3]].index.tolist()
['University of Alaska Anchorage', 'International Academy of Hair Design', 'University of Alabama in Huntsville']
6.3 同时选取DataFrame的行和列
.iloc和.loc可以使用双切片，同时提取行和列：
>>> college = pd.read_csv(
... "data/college.csv", index_col="INSTNM"
... )
>>> college.iloc[:3, :4]
CITY STABBR HBCU MENONLY
INSTNM
Alabama A... Normal AL 1.0 0.0
Universit... Birmingham AL 0.0 0.0
Amridge U... Montgomery AL 0.0 0.0
>>> college.loc[:"Amridge University", :"MENONLY"]
CITY STABBR HBCU MENONLY
INSTNM
Alabama A... Normal AL 1.0 0.0
Universit... Birmingham AL 0.0 0.0
Amridge U... Montgomery AL 0.0 0.0
选取不同两列的所有行：
>>> college.iloc[:, [4, 6]].head()
WOMENONLY SATVRMID
INSTNM
Alabama A & M University 0.0 424.0
University of Alabama at Birmingham 0.0 570.0
Amridge University 0.0 NaN
University of Alabama in Huntsville 0.0 595.0
Alabama State University 0.0 425.0
>>> college.loc[:, ["WOMENONLY", "SATVRMID"]].head()
WOMENONLY SATVRMID
INSTNM
Alabama A & M University 0.0 424.0
University of Alabama at Birmingham 0.0 570.0
Amridge University 0.0 NaN
University of Alabama in Huntsville 0.0 595.0
Alabama State University 0.0 425.0
选取不连续的行和列：
>>> college.iloc[[100, 200], [7, 15]]
SATMTMID UGDS_NHPI
INSTNM
GateWay Community College NaN 0.0029
American Baptist Seminary of the West NaN NaN
>>> rows = [
... "GateWay Community College",
... "American Baptist Seminary of the West",
... ]
>>> columns = ["SATMTMID", "UGDS_NHPI"]
>>> college.loc[rows, columns]
SATMTMID UGDS_NHPI
INSTNM
GateWay Community College NaN 0.0029
American Baptist Seminary of the West NaN NaN
选取一个标量值：
>>> college.iloc[5, -4]
0.401
>>> college.loc["The University of Alabama", "PCTFLOAN"]
0.401
选取单列，对行做切分：
>>> college.iloc[90:80:-2, 5]
INSTNM
Empire Beauty School-Flagstaff 0
Charles of Italy Beauty College 0
Central Arizona College 0
University of Arizona 0
Arizona State University-Tempe 0
Name: RELAFFIL, dtype: int64
>>> start = "Empire Beauty School-Flagstaff"
>>> stop = "Arizona State University-Tempe"
>>> college.loc[start:stop:-2, "RELAFFIL"]
INSTNM
Empire Beauty School-Flagstaff 0
Charles of Italy Beauty College 0
Central Arizona College 0
University of Arizona 0
Arizona State University-Tempe 0
Name: RELAFFIL, dtype: int64
更多
下面两种操作等价：
college.iloc[:10]
college.iloc[:10, :]
6.4 用整数和标签选取数据
先读取数据：
>>> college = pd.read_csv(
... "data/college.csv", index_col="INSTNM"
... )
使用.get_loc找到某一列的序号：
>>> col_start = college.columns.get_loc("UGDS_WHITE")
>>> col_end = college.columns.get_loc("UGDS_UNKN") + 1
>>> col_start, col_end
(10, 19)
用col_start和col_end选取列：
>>> college.iloc[:5, col_start:col_end]
UGDS_WHITE UGDS_BLACK ... UGDS_NRA UGDS_UNKN
INSTNM ...
Alabama A... 0.0333 0.9353 ... 0.0059 0.0138
Universit... 0.5922 0.2600 ... 0.0179 0.0100
Amridge U... 0.2990 0.4192 ... 0.0000 0.2715
Universit... 0.6988 0.1255 ... 0.0332 0.0350
Alabama S... 0.0158 0.9208 ... 0.0243 0.0137
<BLANKLINE>
[5 rows x 9 columns]
更多
行索引切片提取多行多列：
>>> row_start = college.index[10]
>>> row_end = college.index[15]
>>> college.loc[row_start:row_end, "UGDS_WHITE":"UGDS_UNKN"]
UGDS_WHITE UGDS_BLACK ... UGDS_NRA UGDS_UNKN
INSTNM ...
Birmingha... 0.7983 0.1102 ... 0.0000 0.0051
Chattahoo... 0.4661 0.4372 ... 0.0000 0.0139
Concordia... 0.0280 0.8758 ... 0.0466 0.0000
South Uni... 0.3046 0.6054 ... 0.0019 0.0326
Enterpris... 0.6408 0.2435 ... 0.0012 0.0069
James H F... 0.6979 0.2259 ... 0.0007 0.0009
<BLANKLINE>
[6 rows x 9 columns]
6.5 按字母顺序切分
先读取数据：
>>> college = pd.read_csv(
... "data/college.csv", index_col="INSTNM"
... )
尝试选取Sp和Su之间的学校：
>>> college.loc["Sp":"Su"]
Traceback (most recent call last):
...
ValueError: index must be monotonic increasing or decreasing
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
...
KeyError: 'Sp'
报错是因为索引没有排序，对索引做排序：
>>> college = college.sort_index()
重复一开始的操作：
>>> college.loc["Sp":"Su"]
CITY STABBR ... MD_EARN_WNE_P10 GRAD_DEBT_MDN_SUPP
INSTNM ...
Spa Tech ... Ipswich MA ... 21500 6333
Spa Tech ... Plymouth MA ... 21500 6333
Spa Tech ... Westboro MA ... 21500 6333
Spa Tech ... Westbrook ME ... 21500 6333
Spalding ... Louisville KY ... 41700 25000
... ... ... ... ... ...
Studio Ac... Chandler AZ ... NaN 6333
Studio Je... New York NY ... PrivacyS... PrivacyS...
Stylemast... Longview WA ... 17000 13320
Styles an... Selmer TN ... PrivacyS... PrivacyS...
Styletren... Rock Hill SC ... PrivacyS... 9495.5
<BLANKLINE>
[201 rows x 26 columns]
更多
用.is_monotonic_increasing或is_monotonic_decreasing判断索引是否是单调排序的：
>>> college = college.sort_index(ascending=False)
>>> college.index.is_monotonic_decreasing
True
>>> college.loc["E":"B"]
CITY ...
INSTNM ...
Dyersburg State Community College Dyersburg ...
Dutchess Community College Poughkeepsie ...
Dutchess BOCES-Practical Nursing Program Poughkeepsie ...
Durham Technical Community College Durham ...
Durham Beauty Academy Durham ...
... ... ...
Bacone College Muskogee ...
Babson College Wellesley ...
BJ's Beauty & Barber College Auburn ...
BIR Training Center Chicago ...
B M Spurr School of Practical Nursing Glen Dale ...
第01章 Pandas基础
第02章 DataFrame基础运算
第03章 创建和持久化DataFrame
第04章 开始数据分析
第05章 探索性数据分析
第06章 选取数据子集
第07章 过滤行
第08章 索引对齐
