将字符串分成未知数的新列

Question

我有一个如下所示的数据集：

data = tibble(emp = c(1:4), 
                  idstring = c("PER20384|PER49576|PER10837|PER92641",
                          "PER20384|PER49576|PER03875|PER72534", 
                          "PER20384|PER98642|PER17134", 
                          "PER20384|PER98623|PER17134|PER01836|PER1234"))

我想用“|”分隔idstring到单独的列中。但是，我需要最右边的字符（例如“PER92641”）始终位于标记为“Level_1”的列中，最左边的字符根据行中的字符数而变化。

我尝试了一些基本步骤，例如：

data_split = str_split(data$idstring, "\|", simplify = T)
colnames(data_split) = paste0("Level_", ncol(data_split):1)

但是我得到这样的错误输出：

  Level_5    Level_4    Level_3    Level_2    Level_1  
[1,] "PER20384" "PER49576" "PER10837" "PER92641" ""       
[2,] "PER20384" "PER49576" "PER03875" "PER72534" ""       
[3,] "PER20384" "PER98642" "PER17134" ""         ""       
[4,] "PER20384" "PER98623" "PER17134" "PER01836" "PER1234"

它应该是这样的：

   Level_5      Level_4    Level_3    Level_2    Level_1  
[1,]   NA       "PER20384" "PER49576" "PER10837" "PER92641"        
[2,]   NA       "PER20384" "PER49576" "PER03875" "PER72534"        
[3,]   NA         NA       "PER20384" "PER98642" "PER17134"       
[4,] "PER20384" "PER98623" "PER17134" "PER01836" "PER1234"

请注意，理想情况下，我也希望在适用的地方用 NA 代替空白。

我觉得我可以以某种方式颠倒每一行的顺序，然后用 NA 替换空格，然后再添加列名，但我希望在这里可以找到更优雅的解决方案。

Answer 1

这可以通过对 NA 值执行 order 来完成。我们把|处的'idstring'拆分成一个list，得到list个元素的maxlengths（'mx'）。使用它来用 length<- 填充 NA（默认情况下它在末尾而不是开头填充），然后我们 order 基于 NA 元素的向量，rbind list 个元素

lst1 <- strsplit(data$idstring, "|", fixed = TRUE)
mx <- max(lengths(lst1))
out <- do.call(rbind,  lapply(lst1, function(x) {
       length(x) <- mx
        x[order(!is.na(x))]
  }))
colnames(out) <- paste0("Level_", ncol(out):1)

-输出

#    Level_5    Level_4    Level_3    Level_2    Level_1   
#[1,] NA         "PER20384" "PER49576" "PER10837" "PER92641"
#[2,] NA         "PER20384" "PER49576" "PER03875" "PER72534"
#[3,] NA         NA         "PER20384" "PER98642" "PER17134"
#[4,] "PER20384" "PER98623" "PER17134" "PER01836" "PER1234"

或者另一种选择是使用read.table读取列，然后通过重新排列前面的NA元素来修改行值

d1 <- read.table(text = data$idstring, sep="|", header = FALSE, 
    fill = TRUE, na.strings = c(""), col.names = paste0('Level_',  5:1))

d1[1] <- t(apply(d1, 1, function(x) c(x[is.na(x)], x[!is.na(x)])))

将字符串分成未知数的新列

Separate string into new columns of unknown number

r

dplyr

strsplit