按唯一列值分隔数据

Question

Edit to make the problem more clear: I start with a messy CSV file so I need to identify what is the ID and what is the time variable. Or assign ID and time to the data in the variable columns. This question has now been answered below. Here is my data:

col1<-c("ID", "Date","var1","var2","ID","Date","var1","var2","ID","Date","var1","var2")
col2<-c("1","21-11-2015 14:20", "4.8","3.8", "1","21-11-2015 15:30", "3.5","5.9","2","21-11-2015 14:20","3.0","6.7")
df<-cbind(col1,col2)

I tried with dcast() with no luck:

dcast(ID+Date~var1+var2, data = df, value.var = col1 )

I would like the output to be a true long format like this:

ID<-c(1,1,2)
Date<-c("21-11-2015 14:20","21-11-2015 15:30","21-11-2015 14:20")
var1<-c("4.8","3.5","6.7")
var2<-c("3.8","5.9","3.0")
df.clean<-cbind(ID,Date, var1,var2)

I appreciate your help.

Answer 1

Don't think this is a reshape question, you have values in one column and names in other which can be gathered together and given names using setNames

with(df, setNames(data.frame(matrix(col2,
          ncol = length(unique(col1)), byrow = TRUE)), unique(col1)))

#  ID             Date var1 var2
#1  1 21-11-2015 14:20  4.8  3.8
#2  1 21-11-2015 15:30  3.5  5.9
#3  2 21-11-2015 14:20  3.0  6.7

data

col1<-c("ID", "Date","var1","var2","ID","Date","var1","var2","ID",
        "Date","var1","var2")
col2<-c("1","21-11-2015 14:20", "4.8","3.8", "1","21-11-2015 15:30", 
         "3.5","5.9","2","21-11-2015 14:20","3.0","6.7")
df<- data.frame(col1,col2)

Answer 2

This is not a reshape question. Here I supply a simple code on how to do it manually:

Data

col1<-c("ID", 
        "Date","var1","var2","ID","Date","var1","var2","ID","Date","var1","var2")
col2<-c("1","21-11-2015 14:20", "4.8","3.8", "1","21-11-2015 15:30", 
        "3.5","5.9","2","21-11-2015 14:20","3.0","6.7")
df<-data.frame(col1,col2, stringsAsFactors = F)

Code

uniquevars<-unique(col1)
Res<-list()
for(i in 1:length(uniquevars)){
  Res[[uniquevars[i]]]<-df[,"col2"][which(df[,"col1"] ==uniquevars[i])]
}

dfRes <- data.frame(matrix(unlist(Res), ncol=length(Res)),stringsAsFactors=FALSE)
colnames(dfRes)<-uniquevars
dfRes
      ID             Date var1 var2
    1  1 21-11-2015 14:20  4.8  3.8
    2  1 21-11-2015 15:30  3.5  5.9
    3  2 21-11-2015 14:20  3.0  6.7

I hope this code makes you understand the steps to follow on what you are interested in doing.

Cheers !

Answer 3

Here's a tidyverse approach:

library(tidyverse)

df %>%                                # your original (cbind) object
  data.frame() %>%                    # set as dataframe
  group_by(col1) %>%                  # for each col1 value
  mutate(index = row_number()) %>%    # set a row index (useful for reshaping)
  spread(col1, col2) %>%              # reshape
  select(-index)                      # remove index

# # A tibble: 3 x 4
#   Date             ID    var1  var2 
#   <fct>            <fct> <fct> <fct>
# 1 21-11-2015 14:20 1     4.8   3.8  
# 2 21-11-2015 15:30 1     3.5   5.9  
# 3 21-11-2015 14:20 2     3.0   6.7