如何使用 dplyr 执行 separate() 后跟 mutate_each()

Question

我在 sqlite 数据库中有数据，其中包含一个非第一范式的实体。 'sample_attribute' 列中的字符串如下所示：

 isolate: R4166 || age: 43.88 || biomaterial_provider: LIBD || sex: male || tissue: DLPFC || disease: control || race: AA || RIN: 8.7 || Fraction: total || BioSampleModel: Human

我此时的代码：

library(tidyr)
library(dplyr)
library(stringi)



rs.df <- structure(list(run_accession = c("SRR1554537", "SRR2071348"), 
platform_parameters = c("INSTRUMENT_MODEL: Illumina HiSeq 2000", 
"INSTRUMENT_MODEL: Illumina HiSeq 2000"), sample_attribute = c("isolate: R3452 || age: -0.3836 || biomaterial_provider: LIBD || sex: female || tissue: DLPFC || disease: control || race: AA || RIN: 9.6 || Fraction: total || BioSampleModel: Human", "isolate: R3452 || age: -0.3836 || biomaterial_provider: LIBD || sex: female || tissue: DLPFC || disease: control || race: AA || RIN: 9.6 || Fraction: total || BioSampleModel: Human")), .Names = c("run_accession", "platform_parameters", "sample_attribute"
), row.names = c(NA, -2L), class = "data.frame")

coln <- c("isolate", "age", "biomaterial_provider", "sex", "tissue", "disease", "race",
          "RIN", "Fraction", "BioSampleModel")

rs.df <- rs.df %>%
        separate(sample_attribute, coln, sep = "\|\|")

head(rs.df, 1)

中间结果：

       sample_attribute
  run_accession                   platform_parameters         isolate          age
1    SRR1554534 INSTRUMENT_MODEL: Illumina HiSeq 2000 isolate: DLPFC   age: 40.42 
          biomaterial_provider         sex          tissue            disease
1  biomaterial_provider: LIBD   sex: male   tissue: DLPFC   disease: Control 
        race        RIN          Fraction         BioSampleModel
1  race: AA   RIN: 8.4   Fraction: total   BioSampleModel: Human

目前我继续

for (x in coln){
        rs.df[,x] <- stri_replace(rs.df[,x], regex = "^.+:\s*", replacement = "")
}

但这是不灵活的。

有没有办法扩展 dplyr 管道，以便（尽可能）用 %>% 管道中的调用替换 for 循环？

至少，对于 coln 中列的值，从 separate() 调用的结果中删除冒号之前的字符串：

rs.df <- rs.df %>%
        separate(sample_attribute, coln, sep = "\|\|") %>%
        mutate_each(... stri_replace...) #split pairs at ":", remove part before ":"

(这里的for循环解决了我的separating/cleaning字符串向上的问题。但是，SRAdb数据库中可能有更多这样的列，key:valuepairs由“||”分隔。如何以更灵活的方式处理它们？）

Answer 1

在此处查看@docendo discimus 的回答：dplyr certain columns

你的情况

rs.df <- rs.df %>%
    separate(sample_attribute, coln, sep = "\|\|") %>%
    mutate_each_(funs(stri_replace(., regex="^.+:\s*", replacement="")), coln)

如何使用 dplyr 执行 separate() 后跟 mutate_each()

How to perform a separate() followed by a mutate_each() with dplyr

r

dplyr

tidyr