如果一行中存在字符串，return 新列中存在该字符串的列 NAME

Question

我有一个像这样的数据框 text/descriptions:

Year       Descr1              Descr2             Descr3
2017    Abby is cool        Hi, call me 5      Don't call    
1997    Is Doug cool?       Is Gideon OK?     Abby's coming at 5 
2020     Hi i'm Gideon       Hi i'm Abby       OK, ttyl   

My keywords are Doug and Abby. If any of my keywords appear in a row, I want a new column that tells me the column name. See below

```    FY        Descr1            Descr2           Descr3         Found_in
      2017    Abby is cool        Hi call me      Don't call        Descr1
      1997    Is Doug cool?       Abby's coming  Is Gideon OK?      Descr1, Descr2
      2020    Hi i'm Gideon       OK, ttyl       Hi i'm Abby        Descr3    ```

right now I have:

df$Found_in <- ifelse(grepl(keywords,df[,1], ignore.case=TRUE), "Descr1",
           ifelse(grepl(keywords,df[,2], ignore.case=TRUE), "Descr2",
           ifelse(grepl(keywords,df[,3], ignore.case=TRUE), "Descr3", "NA")))    ```

但是我有一个非常大的数据框，所以我想知道是否有一种方法可以搜索我的关键字，如果它们存在于我的数据框中的任何地方，return 它们所在的列名称。

我希望这个问题是有道理的。也请原谅格式，这是我第一次在这里提问

Answer 1

这可以解决问题（我将使用您的第一个数据和解决方案来避免 ifelse）但是必须对字符串进行一些考虑（例如，可以将它们转换得更低，这样更容易实现这种比较）：

#Data
data <- structure(list(Year = c(2017, 1997, 2020), Descr1 = c("Abby is cool", 
"Is Doug cool?", "Hi i'm Gideon"), Descr2 = c("Hi, call me 5", 
"Is Gideon OK?", "Hi i'm Abby"), Descr3 = c("Don't call", "Abby's coming at 5", 
"OK, ttyl")), row.names = c(NA, -3L), class = "data.frame")

#Code
data$Index <- apply(data[,-1],1,function(x) paste(names(x)[which(grepl(pattern = c('Doug|Abby'),x = x))],collapse=','))

它产生：

  Year        Descr1        Descr2             Descr3         Index
1 2017  Abby is cool Hi, call me 5         Don't call        Descr1
2 1997 Is Doug cool? Is Gideon OK? Abby's coming at 5 Descr1,Descr3
3 2020 Hi i'm Gideon   Hi i'm Abby           OK, ttyl        Descr2

如果一行中存在字符串，return 新列中存在该字符串的列 NAME

If string exists in a row, return the column NAME under which the string exists in a new column

r

stringr

grepl