R - 从两个单词的字符串中提取城市名称答案

【问题标题】：R - Extract city name from two-word stringsR - 从两个单词的字符串中提取城市名称
【发布时间】：2018-02-21 20:48:09
【问题描述】：

我正在使用一个包含如下城市名称的数据库：

cities <- c("Fairhope 3NE", "Gadsden 19N", "Selma 13 WNW", "Batesville 8 WNW", 
"Elgin 5 S", "Tucson 11 W", "Williams 35 NNW", "Fallbrook 5 NE", 
"Stovepipe Wells 1 SW", "Cortez 8 SE", "La Junta 17 WSW", "Montrose 11 ENE", 
"Everglades City 5 NE", "Sebring 23 SSE", "Brunswick 23 S", "Newton 11 SW", 
"Newton 8 W", "Watkinsville 5 SSE", "Des Moines 17 E", "Champaign 9 SW", 
"Shabbona 5 NNE", "Bedford 5 WNW", "Manhattan 6 SSW", "Oakley 19 SSW", 
"Bowling Green 21 NNE", "Versailles 3 NNW", "Lafayette 13 SE", 
"Monroe 26 N", "Goodridge 12 NNW", "Chillicothe 22 ENE", "Joplin 24 N", 
"Salem 10 W", "Holly Springs 4 N", "Newton 5 ENE", "Asheville 13 S", 
"Asheville 8 SSW", "Durham 11 W", "Jamestown 38 WSW", "Medora 7 E", 
"Northgate 5 ESE", "Harrison 20 SSE", "Lincoln 11 SW", "Lincoln 8 ENE", 
"Whitman 5 ENE", "Las Cruces 20 N", "Los Alamos 13 W", "Socorro 20 N", 
"Mercury 3 SSW", "Coshocton 8 NNE", "Goodwell 2 E", "Stillwater 2 W", 
"Stillwater 5 WNW", "Coos Bay 8 SW", "Corvallis 10 SSW", "Riley 10 WSW", 
"Blackville 3 W", "McClellanville 7 NE", "Aberdeen 35 WNW", "Buffalo 13 ESE", 
"Pierre 24 S", "Sioux Falls 14 NNE", "Crossville 7 NW", "Austin 33 NW", 
"Bronte 11 NNE", "Edinburg 17 NNE", "Monahans 6 ENE", "Muleshoe 19 S", 
"Palestine 6 WNW", "Panther Junction 2 N", "Necedah 5 WNW")

我只想提取城市名称。以下代码适用于某些情况：

gsub( " .*$", "", cities)

但对于具有两个单词名称的城市（例如 Stovepipe Wells 1 SW 和 La Junta 17 WSW）则失败。

对这些案例有什么解决方案的想法吗？

【问题讨论】：

标签： r regex string gsub

【解决方案1】：

您可以删除所有以数字开头的子字符串：

> sub("\\s*\\d.*", "", cities)
 [1] "Fairhope"         "Gadsden"          "Selma"            "Batesville"       "Elgin"            "Tucson"           "Williams"         "Fallbrook"        "Stovepipe Wells" 
[10] "Cortez"           "La Junta"         "Montrose"         "Everglades City"  "Sebring"          "Brunswick"        "Newton"           "Newton"           "Watkinsville"    
[19] "Des Moines"       "Champaign"        "Shabbona"         "Bedford"          "Manhattan"        "Oakley"           "Bowling Green"    "Versailles"       "Lafayette"       
[28] "Monroe"           "Goodridge"        "Chillicothe"      "Joplin"           "Salem"            "Holly Springs"    "Newton"           "Asheville"        "Asheville"       
[37] "Durham"           "Jamestown"        "Medora"           "Northgate"        "Harrison"         "Lincoln"          "Lincoln"          "Whitman"          "Las Cruces"      
[46] "Los Alamos"       "Socorro"          "Mercury"          "Coshocton"        "Goodwell"         "Stillwater"       "Stillwater"       "Coos Bay"         "Corvallis"       
[55] "Riley"            "Blackville"       "McClellanville"   "Aberdeen"         "Buffalo"          "Pierre"           "Sioux Falls"      "Crossville"       "Austin"          
[64] "Bronte"           "Edinburg"         "Monahans"         "Muleshoe"         "Palestine"        "Panther Junction" "Necedah"         
>

这里，

\\s* - 匹配 0+ 个空格
\\d - 一个数字
.* - 字符串的其余部分。

请参阅regex demo。 sub 只执行一次匹配和替换操作。

【讨论】：

【解决方案2】：

假设每个字符串都以相同的模式结束：

空间
数字
可选空格，
组合ENSW

See code in use here

gsub(" \\d+ ?[ENSW]+$", "", cities)

结果：

 [1] "Fairhope"         "Gadsden"          "Selma"            "Batesville"      
 [5] "Elgin"            "Tucson"           "Williams"         "Fallbrook"       
 [9] "Stovepipe Wells"  "Cortez"           "La Junta"         "Montrose"        
[13] "Everglades City"  "Sebring"          "Brunswick"        "Newton"          
[17] "Newton"           "Watkinsville"     "Des Moines"       "Champaign"       
[21] "Shabbona"         "Bedford"          "Manhattan"        "Oakley"          
[25] "Bowling Green"    "Versailles"       "Lafayette"        "Monroe"          
[29] "Goodridge"        "Chillicothe"      "Joplin"           "Salem"           
[33] "Holly Springs"    "Newton"           "Asheville"        "Asheville"       
[37] "Durham"           "Jamestown"        "Medora"           "Northgate"       
[41] "Harrison"         "Lincoln"          "Lincoln"          "Whitman"         
[45] "Las Cruces"       "Los Alamos"       "Socorro"          "Mercury"         
[49] "Coshocton"        "Goodwell"         "Stillwater"       "Stillwater"      
[53] "Coos Bay"         "Corvallis"        "Riley"            "Blackville"      
[57] "McClellanville"   "Aberdeen"         "Buffalo"          "Pierre"          
[61] "Sioux Falls"      "Crossville"       "Austin"           "Bronte"          
[65] "Edinburg"         "Monahans"         "Muleshoe"         "Palestine"       
[69] "Panther Junction" "Necedah"

【讨论】：

感谢您的回答。它是被接受的，因为a）它有效，b）它是第一个（根据SO报告的时间）。

【解决方案3】：

你可以使用

gsub("(\\D+)\\s+.*", "\\1", cities)

屈服

 [1] "Fairhope"         "Gadsden"          "Selma"            "Batesville"       "Elgin"           
 [6] "Tucson"           "Williams"         "Fallbrook"        "Stovepipe Wells"  "Cortez"          
[11] "La Junta"         "Montrose"         "Everglades City"  "Sebring"          "Brunswick"       
[16] "Newton"           "Newton"           "Watkinsville"     "Des Moines"       "Champaign"       
[21] "Shabbona"         "Bedford"          "Manhattan"        "Oakley"           "Bowling Green"   
[26] "Versailles"       "Lafayette"        "Monroe"           "Goodridge"        "Chillicothe"     
[31] "Joplin"           "Salem"            "Holly Springs"    "Newton"           "Asheville"       
[36] "Asheville"        "Durham"           "Jamestown"        "Medora"           "Northgate"       
[41] "Harrison"         "Lincoln"          "Lincoln"          "Whitman"          "Las Cruces"      
[46] "Los Alamos"       "Socorro"          "Mercury"          "Coshocton"        "Goodwell"        
[51] "Stillwater"       "Stillwater"       "Coos Bay"         "Corvallis"        "Riley"           
[56] "Blackville"       "McClellanville"   "Aberdeen"         "Buffalo"          "Pierre"          
[61] "Sioux Falls"      "Crossville"       "Austin"           "Bronte"           "Edinburg"        
[66] "Monahans"         "Muleshoe"         "Palestine"        "Panther Junction" "Necedah"

解释，这说：

(\\D+) # not a digit, 1+ times
\\s+   # at least one whitespace
.*     # rest of the string

然后将其替换为第一个捕获的组，在本例中为 \\1。

【讨论】：

所以我想“如果一个城市的名称中有一个数字字符怎么办？”。这并不常见（根本不常见），但显然，一些俄罗斯城市曾经有数字Krasnoyask-26。许多closed cities 都这样做了。只是觉得这是一本有趣的书！
@ctwheels：我不知道这个。但是，OP 的城市对我来说确实看起来很美国，不是吗？也可能是加拿大人...
哦，我完全同意，你的回答很好（而且很简单），我显然只是想打破它并发现它，认为它很有趣:)

【解决方案4】：

这应该可行：

gsub( " [1-9].*$", "", cities)

【讨论】：

【解决方案5】：

这个正则表达式应该可以在任何地方使用，包括 R:.+?(?=\ \d)

它包括直到第一个空格后跟一个数字的所有内容。

【讨论】：