Reshaping my data frame and adding a flag

Question

I have a data frame that looks like this:

df <- 
ID   TIME   IPREDP   IPREDM
1     0.5    10        5
1     1.0    15        7
2     0.7    8         2

I want to reshape my data and collapse the IPREDP and IPREDM columns into one column called IPRED and add a flag called DVID for each and get rid of the original IPREDP and IPREDM columns. So the output should be this:

dfout <- 
ID    TIME    DVID    IPRED
1     0.5     1       10
1     0.5     2       5
1     1.0     1       15
1     1.0     2       7
2     0.7     1       8
2     0.7     2       2

How can I achieve this in R in the fastest possible way?

With tidyr and dplyr, `df %>% gather(DVID, IPRED, IPREDP:IPREDM) %>% mutate(DVID = as.integer(factor(DVID, levels = c('IPREDP', 'IPREDM'))))` — alistaire, Aug 09 '16 at 02:03
using library(reshape2), `melt(df,measure.vars =c("IPREDP","IPREDM"), value.name="IPRED")` — dww, Aug 09 '16 at 02:08
In base R - `reshape(df, idvar=c("ID","TIME"), direction="long", varying=3:4, v.names="IPRED", timevar="DVID")` — thelatemail, Aug 09 '16 at 02:20

dww · Accepted Answer · 2016-08-09T03:23:52.533

2

library(reshape2)
dfout <- melt(df,measure=c("IPREDP","IPREDM"), value="IPRED", variable="DVID")
dfout$DVID <- as.integer(dfout$DVID)

edited Aug 09 '16 at 03:23

answered Aug 09 '16 at 02:07

dww

30,425
5
68
111

score 2 · Answer 2 · answered Aug 09 '16 at 02:57

2

   library("data.table") 
   df1 <- melt(df, id = c("ID", "TIME"), measure  = c("IPREDP", "IPREDM"), value.name = "IPRED")
   df1$DVID <- as.integer(df1$variable)
   df1[, variable := NULL]

answered Aug 09 '16 at 02:57

Sathish

12,453
3
41
59

score 1 · Answer 3 · answered Aug 09 '16 at 02:54

1

We can use melt from data.table

library(data.table)
melt(setDT(df), id.var = c("ID", "TIME"), value.name = "IPRED")[
          order(ID)][, DVID := 1:.N,.(ID, variable)][]

answered Aug 09 '16 at 02:54

akrun

874,273
37
540
662

Reshaping my data frame and adding a flag

3 Answers3