r匹配和替换[重复]答案

【问题标题】：r Match and replace [duplicate]r匹配和替换[重复]
【发布时间】：2017-04-02 23:40:41
【问题描述】：

我有一个包含多行和多列的数据集。下面是一些行和列的快照。

  ID  Date        Gender Age  Col1    Col2    Col3    Col4    Col5    Col6    Col7    Col8 
  10  2015-10-14  F      68   345.50  884.2   008.69  202.18  189.8   435.2   084.7   757
  93  2002-07-22  F      87   242.80  710.9   345.50  884.2   008.69  202.18  189.8   435.2  
  14  2004-07-28  M      92   084.7   757     242.80  710.9   427.2   530.10  567.89  227.9
  41  2011-02-24  M      39   714.0   084.7   757     242.80  710.9   427.2   530.10  567.89
  64  2002-03-14  F      39   227.9   714.0   V58.49  906.7   800.35  V88.0   349.31  289.84 
  22  2015-11-21  F      68   324.0   V65.44  411.8   200.41  187.7   E869.3  041.04  170.4
  36  2003-09-17  F      75   389.1   176.3   788.37  E936.3  277.82  812.12  E816.7  663.90
  11  2000-10-07  M      74   716.90  396.3   482.1   E816.7  663.90  716.90  396.3   482.1 
  45  2001-07-14  F      31   614.2   945.44  799.4   864.05  371.31  268     626.2   780.72
  60  1999-02-23  M      45   674     645.2   006.5   V68.2   V67.00  665.24  434.00  914.3

我有另一个数据集，它是一个查找表，其中包含 Col1, Col2, Col3, Col4, Col5, Col6, Col7 和 Col8 中代码的简短描述，如下所示

 Code       Short_Description
 345.50     interStellar
 884.2      indispensable
 008.69     hallucination
 202.18     flow
 189.8      categorizing
 435.2      choppiness
 084.7      chieftain
 757        substantiating
 V58.49     unbridled
 V88.0      polish
 324.0      stumble
 V65.44     hoopster
 411.8      overtrimmed
 E869.3     overbrutalizing
 041.04     choric
 E936.3     busera
 277.82     subdelegating
 E816.7     baton   
 663.90     Space

我的问题是如何将第一个数据集中的代码与第二个查找数据集中的代码进行匹配，并将匹配的代码替换为相应的简短描述？

下面的预期输出显示代码 345.50 匹配并替换为 interStellar ，V58.49 匹配并替换为 unbridled 我希望输出其中所有代码匹配并替换为相应的描述。我知道如何使用 if-then-else 来做到这一点，但这会非常低效，我假设应该有一些简单的方法来做到这一点。任何帮助深表感谢。提前致谢。

  ID  Date        Gender Age  Col1    Col2    Col3    Col4    Col5    Col6    Col7    Col8 
  10  2015-10-14  F      68   interStellar  884.2   008.69  202.18  189.8   435.2   084.7   757
  93  2002-07-22  F      87   242.80  710.9   interStellar  884.2   008.69  202.18  189.8   435.2  
  14  2004-07-28  M      92   084.7   757     242.80  710.9   427.2   530.10  567.89  227.9
  41  2011-02-24  M      39   714.0   084.7   757     242.80  710.9   427.2   530.10  567.89
  64  2002-03-14  F      39   227.9   714.0   unbridled  906.7   800.35  V88.0   349.31  289.84 
  22  2015-11-21  F      68   324.0  hoopster  411.8   200.41  187.7   E869.3  041.04  170.4
  36  2003-09-17  F      75   389.1   176.3   788.37  E936.3  277.82  812.12  baton  663.90
  11  2000-10-07  M      74   716.90  396.3   482.1   baton  663.90  716.90  396.3   482.1 
  45  2001-07-14  F      31   614.2   945.44  799.4   864.05  371.31  268     626.2   780.72
  60  1999-02-23  M      45   674     645.2   006.5   V68.2   V67.00  665.24  434.00  914.3

===================== 本例中使用的可重现数据集=========================

df1 = structure(list(ID = c(10L, 93L, 14L, 41L, 64L, 22L, 36L, 11L, 
45L, 60L), Date = c("10/14/2015", "7/22/2002", "7/28/2004", "2/24/2011", 
"3/14/2002", "11/21/2015", "9/17/2003", "10/7/2000", "7/14/2001", 
"2/23/1999"), Gender = c("F", "F", "M", "M", "F", "F", "F", "M", 
"F", "M"), Age = c(68L, 87L, 92L, 39L, 39L, 68L, 75L, 74L, 31L, 
45L), Col1 = c(345.5, 242.8, 84.7, 714, 227.9, 324, 389.1, 716.9, 
614.2, 674), Col2 = c("884.2", "710.9", "757", "84.7", "714", 
"V65.44", "176.3", "396.3", "945.44", "645.2"), Col3 = c("8.69", 
"345.5", "242.8", "757", "V58.49", "411.8", "788.37", "482.1", 
"799.4", "6.5"), Col4 = c("202.18", "884.2", "710.9", "242.8", 
"906.7", "200.41", "E936.3", "E816.7", "864.05", "V68.2"), Col5 = c("189.8", 
"8.69", "427.2", "710.9", "800.35", "187.7", "277.82", "663.9", 
"371.31", "V67.00"), Col6 = c("435.2", "202.18", "530.1", "427.2", 
"V88.0", "E869.3", "812.12", "716.9", "268", "665.24"), Col7 = c("84.7", 
"189.8", "567.89", "530.1", "349.31", "41.04", "E816.7", "396.3", 
"626.2", "434"), Col8 = c(757, 435.2, 227.9, 567.89, 289.84, 
170.4, 663.9, 482.1, 780.72, 914.3)), class = c("tbl_df", "tbl", 
"data.frame"), row.names = c(NA, -10L), .Names = c("ID", "Date", 
"Gender", "Age", "Col1", "Col2", "Col3", "Col4", "Col5", "Col6", 
"Col7", "Col8"), spec = structure(list(cols = structure(list(
    ID = structure(list(), class = c("collector_integer", "collector"
    )), Date = structure(list(), class = c("collector_character", 
    "collector")), Gender = structure(list(), class = c("collector_character", 
    "collector")), Age = structure(list(), class = c("collector_integer", 
    "collector")), Col1 = structure(list(), class = c("collector_double", 
    "collector")), Col2 = structure(list(), class = c("collector_character", 
    "collector")), Col3 = structure(list(), class = c("collector_character", 
    "collector")), Col4 = structure(list(), class = c("collector_character", 
    "collector")), Col5 = structure(list(), class = c("collector_character", 
    "collector")), Col6 = structure(list(), class = c("collector_character", 
    "collector")), Col7 = structure(list(), class = c("collector_character", 
    "collector")), Col8 = structure(list(), class = c("collector_double", 
    "collector"))), .Names = c("ID", "Date", "Gender", "Age", 
"Col1", "Col2", "Col3", "Col4", "Col5", "Col6", "Col7", "Col8"
)), default = structure(list(), class = c("collector_guess", 
"collector"))), .Names = c("cols", "default"), class = "col_spec"))


lookup_table = structure(list(Code = c("345.5", "884.2", "8.69", "202.18", "189.8", 
"435.2", "84.7", "757", "V58.49", "V88.0", "324", "V65.44", "411.8", 
"E869.3", "41.04", "E936.3", "277.82", "E816.7", "63.9"), Short_Description = c("interStellar", 
"indispensable", "hallucination", "flow", "\tcategorizing", "choppiness", 
"chieftain", "\tsubstantiating", "unbridled", "polish", "stumble", 
"hoopster", "overtrimmed", "overbrutalizing", "choric", "busera", 
"subdelegating", "baton\t", "Space")), class = c("tbl_df", "tbl", 
"data.frame"), row.names = c(NA, -19L), .Names = c("Code", "Short_Description"
), spec = structure(list(cols = structure(list(Code = structure(list(), class = c("collector_character", 
"collector")), Short_Description = structure(list(), class = c("collector_character", 
"collector"))), .Names = c("Code", "Short_Description")), default = structure(list(), class = c("collector_guess", 
"collector"))), .Names = c("cols", "default"), class = "col_spec"))

【问题讨论】：

library(dplyr) ; df1 %>% mutate_at(vars(starts_with('Col')), funs(ifelse(. %in% lookup_table$Code, lookup_table$Short_Description[match(., lookup_table$Code)], .)))
@alistaire，这行得通，如何更改代码的vars(starts_with('Col')) 部分，以便我可以选择以Col 或Secondary 开头的变量
vars(matches('Col|Secondary')) 也许。文档：?dplyr::select_helpers

标签： r replace match

【解决方案1】：

我们可以通过 gather/spread 将“宽”数据集重新整形为“长”格式，最初使用 gather 执行此操作，使用“lookup_table”执行 left_join，mutate 使用“代码”替换'Code' 中带有 'Short_Description' 的元素（其中不是缺失值）和 spread 在 selecting 所需列之后转换为 'wide' 格式（删除 'Short_Description'）。

library(dplyr)
library(tidyr)
gather(df1, Var, Code, matches("Col")) %>% 
      left_join(., lookup_table) %>% 
      mutate(Code = if_else(!is.na(Short_Description), Short_Description, Code)) %>% 
      select(-Short_Description) %>%
      spread(Var, Code)

对于大型数据集，另一个选项是 set 来自 data.table。创建具有“Col”作为子字符串（“nm1”）的列名的数字索引。将“data.frame”转换为“data.table”（setDT(df1)），在指定.SDcols 后循环遍历“nm1”列并将其转换为character（因为预期的输出将具有character来自“Short_Description”列的字符串。使用for 循环并使用set 更改“i”中指定的列和行的“值”（使用match）。

library(data.table)
nm1 <- grep("Col", names(df1))
setDT(df1)[, (nm1) := lapply(.SD, as.character), .SDcols= nm1]
for(j in nm1){
  set(df1, i = which(df1[[j]] %chin% lookup_table$Code), j = j,
    value = lookup_table$Short_Description[match(df1[[j]], lookup_table$Code, nomatch=0)])
 }

df1

【讨论】：

很好的解决方案，但我收到一条错误消息，内存不足，这主要是因为我正在处理一个包含许多列和至少百万行的大型数据集。但是，alistaire，解决方案有效
@KimJenkins 感谢 cmets。我更新了 data.table 选项。你能检查一下是否有效
忽略我之前的评论。最新版本就像黄油上的刀一样。它很光滑 * 100。

【解决方案2】：

您可以为此循环，使用来自qdap 包的lookup

library(qdap)

df1[,13:20] <- NA

for(i in 1:dim(df1)[1]){
  for(j in 1:8){
    df1[i,j+12] <- lookup(df1[i,j+4], lookup_table)
  }
}

head(df1)

  ID       Date Gender Age  Col1   Col2   Col3   Col4   Col5   Col6   Col7   Col8          V13              V14              V15           V16            V17             V18            V19              V20
1 10 10/14/2015      F  68 345.5  884.2   8.69 202.18  189.8  435.2   84.7    757 interStellar    indispensable    hallucination          flow \tcategorizing      choppiness      chieftain \tsubstantiating
2 93  7/22/2002      F  87 242.8  710.9  345.5  884.2   8.69 202.18  189.8  435.2         <NA>             <NA>     interStellar indispensable  hallucination            flow \tcategorizing       choppiness
3 14  7/28/2004      M  92  84.7    757  242.8  710.9  427.2  530.1 567.89  227.9    chieftain \tsubstantiating             <NA>          <NA>           <NA>            <NA>           <NA>             <NA>
4 41  2/24/2011      M  39   714   84.7    757  242.8  710.9  427.2  530.1 567.89         <NA>        chieftain \tsubstantiating          <NA>           <NA>            <NA>           <NA>             <NA>
5 64  3/14/2002      F  39 227.9    714 V58.49  906.7 800.35  V88.0 349.31 289.84         <NA>             <NA>        unbridled          <NA>           <NA>          polish           <NA>             <NA>
6 22 11/21/2015      F  68   324 V65.44  411.8 200.41  187.7 E869.3  41.04  170.4      stumble         hoopster      overtrimmed          <NA>           <NA> overbrutalizing         choric             <NA>

【讨论】：

很欣赏这个解决方案，但我什至不会尝试这个选项，因为就像我之前提到的那样，我的数据集中有数百列和数百万行，如果我尝试你的方法，看起来我会添加两倍的列数，这将导致内存问题。