cSplit Coerces不必要的NA Row

Question

我有一个大型数据集，其中一小部分样本看起来像下面的4 x 5。我尝试使用变量library(splitstackshape) dt <- tibble( a = c("Quartz | White Spirit | Wildfire", "Quiet Riot", "Race Against Time", "Down | Heart Lane | X | Breaking H"), b = c("Muthas Pride", "Killer Girls / Slick Black Cadillac", "Demo 1980", "Life 55"), c = c("Split", "Single", "Demo", "Split"), d = c("Birmingham, England | Hartlepool, England | Sheffield, South Yorkshire, England", "Los Angeles, California", "Nottingham, England", "Liverpool | Beijing | | NYC"), e = c("wf | ef | ff", "g", "f", "cf | af | df | rf") ) dt.s <- subset(dt, c == "Split") dt.split <- cSplit(dt.s, c("a", "d", "e"), c("|", "|", "|"), "long") dt.split 将多个分隔列拆分为唯一行，如下所示：

             a            b     c                                   d  e
1:       Quartz Muthas Pride Split                 Birmingham, England wf
2: White Spirit Muthas Pride Split                 Hartlepool, England ef
3:     Wildfire Muthas Pride Split Sheffield, South Yorkshire, England ff
4:           NA Muthas Pride Split                                  NA NA
5:         Down      Life 55 Split                           Liverpool cf
6:   Heart Lane      Life 55 Split                             Beijing af
7:            X      Life 55 Split                                     df
8:   Breaking H      Life 55 Split                                 NYC rf

然而，这会强制增加一行NAs，如第4行所示：

cSplit

如果我只拆分两列，这不是问题。如何让它不生成NA行？而且，有没有办法让c无法按{{1}}进行分组？

Answer 1

当我们使用tibble时，我们可以使用separate_rows，但不会提供NA行

library(tidyr)
separate_rows(dt.s, c('a', "d", "e"), sep="\\s*\\|\\s*") %>%
         select_at(names(dt.s))
# A tibble: 7 x 5
#             a            b     c                                   d     e
#         <chr>        <chr> <chr>                               <chr> <chr>
#1       Quartz Muthas Pride Split                 Birmingham, England    wf
#2 White Spirit Muthas Pride Split                 Hartlepool, England    ef
#3     Wildfire Muthas Pride Split Sheffield, South Yorkshire, England    ff
#4         Down      Life 55 Split                           Liverpool    cf
#5   Heart Lane      Life 55 Split                             Beijing    af
#6            X      Life 55 Split                                        df
#7   Breaking H      Life 55 Split                                 NYC    rf

关于为什么cSplit给出额外的NA行，最好检查“＆＃39;宽”中的输出。格式

cSplit(dt.s, c("a", "d", "e"), "|")
#              b     c    a_1          a_2      a_3        a_4                 d_1                 d_2                                 d_3 d_4 e_1 e_2 e_3 e_4
#1: Muthas Pride Split Quartz White Spirit Wildfire         NA Birmingham, England Hartlepool, England Sheffield, South Yorkshire, England  NA  wf  ef  ff  NA
#2:      Life 55 Split   Down   Heart Lane        X Breaking H           Liverpool             Beijing                                     NYC  cf  af  df  rf

在这里，我们发现对于第二行，分隔符数|为4，为第一行创建NA，因为＆＃39; a列只有3个分隔符。所以，当我们使用＆＃39; long＆＃39;格式，此NA行传播。这可能是一个错误。

Answer 2

尝试将makeEqual = FALSE添加到cSplit来电：

cSplit(dt.s, c("a", "d", "e"), "|", "long", makeEqual = FALSE)
##               a            b     c                                   d  e
## 1:       Quartz Muthas Pride Split                 Birmingham, England wf
## 2: White Spirit Muthas Pride Split                 Hartlepool, England ef
## 3:     Wildfire Muthas Pride Split Sheffield, South Yorkshire, England ff
## 4:         Down      Life 55 Split                           Liverpool cf
## 5:   Heart Lane      Life 55 Split                             Beijing af
## 6:            X      Life 55 Split                                     df
## 7:   Breaking H      Life 55 Split                                 NYC rf

此外，由于您已经使用了＆＃34; tidyverse＆＃34;中的软件包，因此您可以进行分组，如下所示：

dt %>% 
  filter(c == "Split") %>% 
  cSplit(c("a", "d", "e"), "|", "long", makeEqual = FALSE)