CODEWITHSUNDEEP
rappend
Jan
Jan

Reputation: 1867

R append columns in different ways

I have a dataset with multiple columns:

Value1 Value3 Annotation Value4 Value2
1        4        s        9       4
2        5        t        0       4
3        6        q        4       4

The order of the values and the position of Annotation is unknown.

Now I'm trying to append multiple columns and try to keep the Annotation correct. The result should be a 2 column matrix

eg. when I try to append Value1 and Value3:

Annotation Valuelist
s             1
t             2
q             3
s             4
t             5
q             6

I found the method append() but I can't figure out how to make sure the annotation is correct.

Upvotes: 0

Views: 3032

Answers (2)

A5C1D2H2I1M1N2O1R2T1
A5C1D2H2I1M1N2O1R2T1

Reputation: 193527

The stack function might come in handy for what you want to do in base R:

cbind(mydf["Annotation"], stack(mydf[c("Value1", "Value3")])["values"])
#   Annotation values
# 1          s      1
# 2          t      2
# 3          q      3
# 4          s      4
# 5          t      5
# 6          q      6

stack normally creates two columns, "values" and "ind" (in that order), so we select just the first to match what you describe.

A similar approach can be taken with melt from "reshape2":

library(reshape2)
melt(mydf, id.vars="Annotation", 
     measure.vars=c("Value1", "Value3"))[c("Annotation", "value")]
#   Annotation value
# 1          s     1
# 2          t     2
# 3          q     3
# 4          s     4
# 5          t     5
# 6          q     6

Upvotes: 2

josliber
josliber

Reputation: 44330

You could construct this with something like:

columns <- c("Value1", "Value3")
data.frame(Annotation=rep(dat$Annotation, length(columns)),
           Valuelist=as.vector(as.matrix(dat[columns])))
#   Annotation Valuelist
# 1          s         1
# 2          t         2
# 3          q         3
# 4          s         4
# 5          t         5
# 6          q         6

Upvotes: 1

Related Questions