pandas文本处理

  1 import pandas as pd
  2 import numpy as np
  3 
  4 s = pd.series(['a', 'b', 'c', 'bbhello', '123', np.nan, 'hj'])
  5 df = pd.dataframe({'key1': list('abcdef'),
  6                    'key2': ['hee', 'fv', 'w', 'hija', '123', np.nan]})
  7 print(s)
  8 print('-'*8)
  9 print(df)
 10 print('-'*8)
 11 '''
 12 0          a
 13 1          b
 14 2          c
 15 3    bbhello
 16 4        123
 17 5        nan
 18 6         hj
 19 dtype: object
 20 --------
 21   key1  key2
 22 0    a   hee
 23 1    b    fv
 24 2    c     w
 25 3    d  hija
 26 4    e   123
 27 5    f   nan
 28 --------
 29 '''
 30 # 直接通过.str调用字符串方法，可以对series、dataframe使用，自动过滤nan值
 31 print(s.str.count('b'))
 32 '''
 33 0    0.0
 34 1    1.0
 35 2    0.0
 36 3    2.0
 37 4    0.0
 38 5    nan
 39 6    0.0
 40 dtype: float64
 41 '''
 42 print(df['key2'].str.upper())
 43 '''
 44 0     hee
 45 1      fv
 46 2       w
 47 3    hija
 48 4     123
 49 5     nan
 50 name: key2, dtype: object
 51 '''
 52 # 将所有的列名改为大写
 53 df.columns = df.columns.str.upper()
 54 print(df)
 55 '''
 56   key1  key2
 57 0    a   hee
 58 1    b    fv
 59 2    c     w
 60 3    d  hija
 61 4    e   123
 62 5    f   nan
 63 '''
 64 # 字符串常用方法 --lower，upper，len，starswith，endswith
 65 
 66 print('小写，lower()',s.str.lower())
 67 print('大写，upper()',s.str.upper())
 68 print('长度，len()',s.str.len())
 69 print('判断起始是否为b，startswith()',s.str.startswith('b'))
 70 print('判断结束是否为"o"，endswith()',s.str.endswith('o'))
 71 '''
 72 小写，lower() 0          a
 73 1          b
 74 2          c
 75 3    bbhello
 76 4        123
 77 5        nan
 78 6         hj
 79 dtype: object
 80 大写，upper() 0          a
 81 1          b
 82 2          c
 83 3    bbhello
 84 4        123
 85 5        nan
 86 6         hj
 87 dtype: object
 88 长度，len() 0    1.0
 89 1    1.0
 90 2    1.0
 91 3    7.0
 92 4    3.0
 93 5    nan
 94 6    2.0
 95 dtype: float64
 96 判断起始是否为b，startswith() 0    false
 97 1     true
 98 2    false
 99 3     true
100 4    false
101 5      nan
102 6    false
103 dtype: object
104 判断结束是否为"o"，endswith() 0    false
105 1    false
106 2    false
107 3     true
108 4    false
109 5      nan
110 6    false
111 dtype: object
112 '''
113 # 字符串常用方法 --strip
114 
115 s2 = pd.series([' jack', 'jill ', ' jesse  '])
116 df2 = pd.dataframe(np.random.randn(3, 2), columns=[' a ', ' b'], index=range(3))
117 print(s2)
118 print('-'*8)
119 print(df2)
120 print('-'*8)
121 '''
122 0        jack
123 1       jill 
124 2     jesse  
125 dtype: object
126 --------
127          a          b
128 0 -0.333042 -0.467830
129 1  0.605179 -0.658910
130 2 -0.490881 -0.639754
131 --------
132 '''
133 print(s2.str.strip())
134 print('-'*8)
135 print(s2.str.lstrip())
136 print('-'*8)
137 print(s2.str.rstrip())
138 '''
139 0     jack
140 1     jill
141 2    jesse
142 dtype: object
143 --------
144 0       jack
145 1      jill 
146 2    jesse  
147 dtype: object
148 --------
149 0      jack
150 1      jill
151 2     jesse
152 dtype: object
153 '''
154 df2.columns = df2.columns.str.strip()
155 print(df2)
156 '''
157           a         b
158 0 -0.801508  1.650113
159 1 -0.669556 -1.195999
160 2  0.277338 -0.727100
161 
162 '''
163 
164 # 字符串常用方法  -- replace()
165 df3 = pd.dataframe(np.random.randn(3, 2), columns=[' a a', ' b  b'], index=range(3))
166 df3.columns = df3.columns.str.replace(' ', '-', n=2)
167 print(df3)
168 '''
169        -a-a     -b- b
170 0 -1.225938  0.296270
171 1  0.769037  2.794032
172 2 -1.686818  0.109314
173 '''
174 # 字符串常用方法 -- spilt、rsplit
175 s4 = pd.series(['a,b,c', '1,2,3', ['a,,,c'], np.nan])
176 print(s4)
177 print(s4.str.split(','))
178 '''
179 0      a,b,c
180 1      1,2,3
181 2    [a,,,c]
182 3        nan
183 dtype: object
184 0    [a, b, c]
185 1    [1, 2, 3]
186 2          nan
187 3          nan
188 dtype: object
189 '''
190 # 直接索引得到一个list
191 # 可以使用get或[]符号访问拆散列表中的元素
192 print(s4.str.split(',').str[0])
193 print(s4.str.split(',').str.get(0))
194 '''
195 0      a
196 1      1
197 2    nan
198 3    nan
199 dtype: object
200 0      a
201 1      1
202 2    nan
203 3    nan
204 dtype: object
205 '''
206 
207 # 可以使用expand可以轻松扩展此操作以返回dataframe
208 # n 参数限制分割数
209 print(s4.str.split(','))
210 print('-' * 8)
211 print(s4.str.split(',', expand=true))
212 '''
213 0    [a, b, c]
214 1    [1, 2, 3]
215 2          nan
216 3          nan
217 dtype: object
218 --------
219      0    1    2
220 0    a    b    c
221 1    1    2    3
222 2  nan  nan  nan
223 3  nan  nan  nan
224 '''
225 print(s4.str.split(',', expand=true, n=1))
226 '''
227      0    1
228 0    a  b,c
229 1    1  2,3
230 2  nan  nan
231 3  nan  nan
232 '''
233 # rsplit类似于split，反向工作，即从字符串的末尾到字符串的开头
234 print(s4.str.split(',', expand=true, n=1))
235 print('-' * 8)
236 print(s4.str.rsplit(',', expand=true, n=1))
237 '''
238      0    1
239 0    a  b,c
240 1    1  2,3
241 2  nan  nan
242 3  nan  nan
243 --------
244      0    1
245 0  a,b    c
246 1  1,2    3
247 2  nan  nan
248 3  nan  nan
249 '''
250 
251 df4 = pd.dataframe({'key1': ['a,b,c', '1,2,3', [':,,, ']],
252                     'key2': ['a-b-c', '1-2-3', [':-.- ']]})
253 print(df4)
254 print('-'*8)
255 print(df4['key2'].str.split('-'))
256 '''
257       key1     key2
258 0    a,b,c    a-b-c
259 1    1,2,3    1-2-3
260 2  [:,,, ]  [:-.- ]
261 --------
262 0    [a, b, c]
263 1    [1, 2, 3]
264 2          nan
265 name: key2, dtype: object
266 '''
267 # 通过索引获取分割后的元素
268 df4['k201'] = df4['key2'].str.split('-').str[0]
269 df4['k202'] = df4['key2'].str.split('-').str[1]
270 df4['k203'] = df4['key2'].str.split('-').str[2]
271 print(df4)
272 '''
273       key1     key2 k201 k202 k203
274 0    a,b,c    a-b-c    a    b    c
275 1    1,2,3    1-2-3    1    2    3
276 2  [:,,, ]  [:-.- ]  nan  nan  nan
277 '''