Question

最近我写了一段不符合预期的代码。在两个数据帧df1，df2的循环中，我试图在矩阵中收集数据＆＃34; a＆＃34; ...：

df1=data.frame(x=c(1,2),y=c(10,20))
df2=data.frame(x=c(4,5),y=c(40,50))
dlist <- list(df1,df2)

a <- c(0,"...creation...")
a <- rbind(a, c(0,"rbind test OK"))

lapply(dlist, function(d) {
   print(paste("x",d$x[1]))
   a <- rbind(a, c(d$x[1],"data copied"))
   })

a <- rbind(a, c(0,"rbind test 2 OK"))

检查＆＃34; a＆＃34;执行这些行后产生

a
  [,1] [,2]             
a "0"  "...creation..." 
  "0"  "rbind test OK"  
  "0"  "rbind test 2 OK"

也就是说，lapply-loop中的rbind-statement没有被执行。

预期输出为：

a 
  [,1] [,2] 
a "0"  "...creation..." 
  "0"  "rbind test OK" 
  "1"  "data copied" 
  "4"  "data copied" 
  "0"  "rbind test 2 OK"

为什么？

Answer 1

您可以使用运算符<<-修改环境外部环境，但这是一个非常不合适的命令式编程反射：

lapply(dlist, function(d) {
   print(paste("x",d$x[1]))
   a <<- rbind(a, c(d$x[1],"data copied"))
})

正确的方法是：

res = do.call(rbind, lapply(dlist, function(d) c(d$x[1],"data copied")))

a = rbind(a, res, c(0,"rbind test 2 OK"))

#  [,1] [,2]             
#a "0"  "...creation..."  
#  "0"  "rbind test OK"  
#  "1"  "data copied"    
#  "4"  "data copied"    
#  "0"  "rbind test 2 OK"

lapply（list）和＆＃34; for＆＃34; -loop之间的差异

1 个答案: