时间戳

首页 > 解决方案 > 匹配旧列名并替换为来自另一个数据帧的新列名（+ 交互）

r - 匹配旧列名并替换为来自另一个数据帧的新列名（+ 交互）

问题描述

我有一个数据框 ( df.subset)，它混合了原始变量和相互之间的一些交互 (cfr. labs)。我想使用外部数据框 ( df_names) 通过匹配旧名称并用新名称替换它们来重命名列（警告：交互也必须重命名！）。

set.seed(12345)

x1<- rnorm(100, 1000, 10)
x2<- rnorm(100, 200,1)
x3<- rnorm(100, 10000,400)
x4<- rnorm(100, 2405,28)

# starting dataframe
df<-data.frame(cbind(x1,x2,x3,x4))

# i'm interested only in the variables and interactions contained in labs
labs<-c("x1", "x2", "x3", "x4", "x1*x2", "x1*x4", "x2*x3")
str(labs)

# all variables + all interactions
res <- cbind(df, do.call(cbind,combn(colnames(df), 2, 
                                        FUN= function(x) list(df[x[1]]*df[x[2]]))))

colnames(res)[-(seq_len(ncol(df)))] <-  combn(colnames(df), 2, 
                                               FUN = paste, collapse="*")

res

# final dataset
df.subset <- res[, labs]
df.subset

# use this df for new column names
df_names <- data.frame(old = c("x1","x2","x3","x4","x5")
                       , new = c("tv", "radio", "outdoor", "cinema","something_else"))

df_names

我试过mapvalues但没有成功

plyr::mapvalues(names(df.subset), from = df_names$old, to = df_names$new)

输出应如下所示：

> colnames(df.subset)
[1] "tv"    "radio"    "outdoor"    "cinema"    "tv*radio" "tv*cinema" "radio*outdoor"

任何帮助将不胜感激。

谢谢

标签： r

解决方案

我认为一个简单的循环是最简单的方法：

names(df.subset)
#> [1] "x1"    "x2"    "x3"    "x4"    "x1*x2" "x1*x4" "x2*x3"

for(i in seq(nrow(df_names))){
  names(df.subset) <- gsub(df_names$old[i], df_names$new[i], names(df.subset))
}

names(df.subset)
#> [1] "tv"            "radio"         "outdoor"       "cinema"       
#> [5] "tv*radio"      "tv*cinema"     "radio*outdoor"

由reprex 包于 2020-12-08 创建(v0.3.0)

推荐阅读