字符向量列表的混合数据帧到统一的数据帧

Question

我想分解字符串作为列使用stringr包。

> df <- dput(head(facs,3)) 
structure(list(geo_accession = structure(1:3, .Names = c("V2", 
"V3", "V4"), .Label = c("GSM1494875", "GSM1494877", "GSM1494879", 
"GSM1494881", "GSM1494883", "GSM1494885", "GSM1494887", "GSM1494889", 
"GSM1494891", "GSM1494893", "GSM1494895", "GSM1494897", "GSM1494899", 
"GSM1494901", "GSM1494903", "GSM1494906", "GSM1494908", "GSM1494910", 
"GSM1494912", "GSM1494914", "GSM1494917", "GSM1494919", "GSM1494921", 
"GSM1494923"), class = "factor"), title = structure(1:3, .Names = c("V2", 
"V3", "V4"), .Label = c("2818 - Akt treated TIL - repeat 1 - mAdbID:125971", 
"2818 - Akt treated TIL - repeat 2 - mAdbID:125972", "2818 - Akt treated TIL - repeat 3 - mAdbID:125973", 
"2818 - Akt treated TIL - repeat 4 - mAdbID:125974", "2818 - Untreated TIL - repeat 1 - mAdbID:125975", 
"2818 - Untreated TIL - repeat 2 - mAdbID:125976", "2818 - Untreated TIL - repeat 3 - mAdbID:125977", 
"2818 - Untreated TIL - repeat 4 - mAdbID:125978", "3289 - Akt treated TIL - repeat 1 - mAdbID:125979", 
"3289 - Akt treated TIL - repeat 2 - mAdbID:125980", "3289 - Akt treated TIL - repeat 3 - mAdbID:125981", 
"3289 - Akt treated TIL - repeat 4 - mAdbID:125982", "3289 - Untreated TIL - repeat 1 - mAdbID:125983", 
"3289 - Untreated TIL - repeat 2 - mAdbID:125984", "3289 - Untreated TIL - repeat 3 - mAdbID:125985", 
"3289 - Untreated TIL - repeat 4 - mAdbID:125986", "3784 - Akt treated TIL - repeat 1 - mAdbID:125987", 
"3784 - Akt treated TIL - repeat 2 - mAdbID:125988", "3784 - Akt treated TIL - repeat 3 - mAdbID:125989", 
"3784 - Akt treated TIL - repeat 4 - mAdbID:125990", "3784 - Untreated TIL - repeat 1 - mAdbID:125991", 
"3784 - Untreated TIL - repeat 2 - mAdbID:125992", "3784 - Untreated TIL - repeat 3 - mAdbID:125993", 
"3784 - Untreated TIL - repeat 4 - mAdbID:125994"), class = "factor")), .Names = c("geo_accession", 
"title"), row.names = c("GSM1494875", "GSM1494877", "GSM1494879" 
), class = "data.frame") 

现在它是一个2列数据框。

> sapply(df, class) 
geo_accession   title 
"factor"  "factor" 

如果我使用str_split_fixed，

> df$title = str_split_fixed(df[,"title"], " - ", 4) 
> df 
      geo_accession title.1   title.2 title.3  title.4 
GSM1494875 GSM1494875 2818 Akt treated TIL repeat 1 mAdbID:125971 
GSM1494877 GSM1494877 2818 Akt treated TIL repeat 2 mAdbID:125972 

> sapply(df, class) 
geo_accession   title 
"factor"  "matrix" 

我想看到的是一个5列数据帧，而不是2列混合数据帧（因子，矩阵）。

> df 
     geo_accession title1   title2 title3  title4 
GSM1494875 GSM1494875 2818 Akt treated TIL repeat 1 mAdbID:125971 
GSM1494877 GSM1494877 2818 Akt treated TIL repeat 2 mAdbID:125972 

> sapply(df, class) 
geo_accession  title.1  title.2  title.3  title.4 
"factor"  "factor"  "factor"  "factor"  "factor" 

我不知道如何清理混合数据帧。我如何将其转换为5列数据框？

Answer 1

您可以使用separate从tidyr包：

library(tidyr) 
df %>% separate(title, paste0('title',1:4) ,' - ') 

      geo_accession title1   title2 title3  title4 
1 GSM1494875 GSM1494875 2818 Akt treated TIL repeat 1 mAdbID:125971 
2 GSM1494877 GSM1494877 2818 Akt treated TIL repeat 2 mAdbID:125972 

Answer 2

str_split_fixed的输出是list的vectors。将list指定给单列是导致问题的原因。一种方法是将其分配到多个列或以其他方式，我们可以do.call(data.frame解决这个问题，即

facs1 <- do.call(data.frame, facs) 

---

或者我们可以前后-，然后cbind去掉空格后使用base R方法与read.csv/read.table与第一列

facs2 <- cbind(facs[1], read.csv(text=gsub("\\s+-\\s+", "-", facs$title), 
      sep="-", header=FALSE, col.names = paste0("title.", 1:4), stringsAsFactors=FALSE)) 
facs2 
#   geo_accession title.1   title.2 title.3  title.4 
#GSM1494875 GSM1494875 2818 Akt treated TIL repeat 1 mAdbID:125971 
#GSM1494877 GSM1494877 2818 Akt treated TIL repeat 2 mAdbID:125972 

str(facs2) 
#'data.frame': 2 obs. of 5 variables: 
#$ geo_accession: chr "GSM1494875" "GSM1494877" 
#$ title.1  : int 2818 2818 
#$ title.2  : chr "Akt treated TIL" "Akt treated TIL" 
#$ title.3  : chr "repeat 1" "repeat 2" 
#$ title.4  : chr "mAdbID:125971" "mAdbID:125972" 

---

或压缩选项是cSplitsplitstackshape

library(splitstackshape) 
cSplit(facs, "title", " - ")