A data frame contains several TIMESTAMP Columns.
# Create simple example
df <- cbind.data.frame(A=c("A","B","C")
,B=c(1,2,3)
,Timestamp_1=c(as.POSIXct(NA),as.POSIXct("2018-05-04 00:19:41"),as.POSIXct("2018-07-31 22:09:10"))
,Timestamp_2=c(as.POSIXct("2018-05-04 00:18:45"),as.POSIXct("2018-05-05 00:18:43"),as.POSIXct("2018-06-05 00:00:01"))
,Timestamp_3=c(as.POSIXct("2018-05-04 00:19:13"),as.POSIXct("2018-05-05 00:17:00"),as.POSIXct("2018-05-06 00:18:41"))
,C=c("Dog","Cat","Mouse")
)
df
A B Timestamp_1 Timestamp_2 Timestamp_3 C
1 A 1 <NA> 2018-05-04 00:18:45 2018-05-04 00:19:13 Dog
2 B 2 2018-05-04 00:19:41 2018-05-05 00:18:43 2018-05-05 00:17:00 Cat
3 C 3 2018-07-31 22:09:10 2018-06-05 00:00:01 2018-05-06 00:18:41 Mouse
The result of merging the TIMESTAMP Columns needs to look like
df_Result
A B Timestamp_ALL C
1 A 1 <NA> Dog
2 A 1 2018-05-04 00:18:45 Dog
3 A 1 2018-05-04 00:19:13 Dog
4 B 2 2018-05-04 00:19:41 Cat
5 B 2 2018-05-05 00:17:00 Cat
6 B 2 2018-05-05 00:18:43 Cat
7 C 3 2018-05-06 00:18:41 Mouse
8 C 3 2018-06-05 00:00:01 Mouse
9 C 3 2018-07-31 22:09:10 Mouse
How to do this in a R-elegant & efficient way? Thanks a million for any advice and idea!
Here is a tidyverse solution using tidyr::gather
library(tidyverse)
df %>%
gather(key, Timestamp_ALL, -A, -B, -C) %>%
select(A, B, Timestamp_ALL, C, -key) %>%
arrange(A, B)
# A B Timestamp_ALL C
#1 A 1 <NA> Dog
#2 A 1 2018-05-04 00:18:45 Dog
#3 A 1 2018-05-04 00:19:13 Dog
#4 B 2 2018-05-04 00:19:41 Cat
#5 B 2 2018-05-05 00:18:43 Cat
#6 B 2 2018-05-05 00:17:00 Cat
#7 C 3 2018-07-31 22:09:10 Mouse
#8 C 3 2018-06-05 00:00:01 Mouse
#9 C 3 2018-05-06 00:18:41 Mouse
Explanation: gather converts the data.frame from wide to long, the rest is just selecting relevant columns using select, and ordering entries by A then B with arrange.