如何在 R 中解析多个分隔数据中的列/值答案

【问题标题】：How to parse Column/Value in multiple delimited data in R如何在 R 中解析多个分隔数据中的列/值
【发布时间】：2020-08-28 16:03:51
【问题描述】：

我收到了一个 0365 的奇怪文件，它似乎由 : 和 , 用引号分隔。我想将它们放入单独的列和值中。

以下示例：

CreationDate UserID  AuditData 
2020-05-04   User1   {"Id":"4ccd2","RecordType":20,"CreationTime":"2020-05-04T10:24:44"} 
2020-04-14   User2   {"Id":"4def5","RecordType":18,"CreationTime":"2020-04-14T10:24:44"} 
2020-03-29   User3   {"Id":"4zxc2","RecordType":07,"CreationTime":"2020-03-29T10:24:44"}

目标：将 AuditData 列分解为： 1）身份和价值 2）记录类型和值 3) 创建时间和价值

等等等等

我一直在尝试使用 separate() 做一些事情，但到目前为止都没有成功。谢谢！

【问题讨论】：

标签： r dplyr csv

【解决方案1】：

这是使用separate 的tidyverse 解决方案。

#Your data
df<-read.csv(text = 'CreationDate UserID AuditData
2020-05-04 User1 {"Id":"4ccd2","RecordType":20,"CreationTime":"2020-05-04T10:24:44"}
2020-04-14 User2 {"Id":"4def5","RecordType":18,"CreationTime":"2020-04-14T10:24:44"}
2020-03-29 User3 {"Id":"4zxc2","RecordType":07,"CreationTime":"2020-03-29T10:24:44"}',
         sep = " ")

library(tidyverse)
df %>%
   # remove keys using gsub
   mutate_at(vars(AuditData), function(x) gsub("\\{|\\}","",x)) %>%
   # separate using the colon or comma (however this separates also the time values)
   separate(col = AuditData, 
            # Define the new column names
            into = c("Id","Idvalue","RecordType","RecordTypevalue","CreationTime","temp","time1","time2"),
            # Use : or , as separators
            sep = "\\:|\\,") %>%
   # Use paste to reconstruct the time values
   mutate(CreationTimevalue = paste(temp,time1,time2, sep = ":")) %>%
   # Eliminate unused columns: temp, time1 and time2 
   select(-c(temp,time1,time2))

# CreationDate UserID Id Idvalue RecordType RecordTypevalue CreationTime   CreationTimevalue
# 1   2020-05-04  User1 Id   4ccd2 RecordType              20 CreationTime 2020-05-04T10:24:44
# 2   2020-04-14  User2 Id   4def5 RecordType              18 CreationTime 2020-04-14T10:24:44
# 3   2020-03-29  User3 Id   4zxc2 RecordType              07 CreationTime 2020-03-29T10:24:44

【讨论】：