How to reshape any data-frame to a 2-columns data-frame, with the (repeated) column names in the first and the corresponding values in the second?

Question

How can one reshape data-frame like this:

x1  x2  x3  x4
1   1   5   6
1   2   3   5
2   1   6   1
2   2   2   4

to:

x1  1   
x1  1
x1  2
x1  2
x2  1
.
.
.

I have read many posts (like this one) but I could not figure this out.

`data.frame(nm = rep(names(df), each = nrow(df)), value = unlist(df), row.names = NULL)` — d.b, Oct 31 '17 at 16:06
This questions is not useful. You said "I have read many posts but I could not figure this out" but a simple "R + reshape" search on this website brings on this page: https://stackoverflow.com/questions/5890584/how-to-reshape-data-from-long-to-wide-format, which I assume contains your answer. (And if it does not, you have to explain why.) — Arthur, Oct 31 '17 at 16:21
Possible duplicate of [How to reshape data from long to wide format?](https://stackoverflow.com/questions/5890584/how-to-reshape-data-from-long-to-wide-format) — Arthur, Oct 31 '17 at 16:25
@Arthur this is not unstacking problem. thanks for that post but not really useful — nik, Oct 31 '17 at 16:36
OK, and sorry for being rude. Maybe the title of your question is misleading... — Arthur, Oct 31 '17 at 16:37
@Arthur you can always modify the title and the question. I am always happy to learn and I appreciate the time you have put and looked at my question. Thanks — nik, Oct 31 '17 at 16:40

BENY · Accepted Answer · 2017-10-31T16:07:46.253

2

By using stack

stack(df)
   values ind
1       1  x1
2       1  x1
3       2  x1
4       2  x1
5       1  x2
6       2  x2
7       1  x2
8       2  x2
9       5  x3
10      3  x3
11      6  x3
12      2  x3
13      6  x4
14      5  x4
15      1  x4
16      4  x4

For plot

df$ind=as.character(df$ind)
boxplot(values~ind,data=df)

edited Oct 31 '17 at 16:07

answered Oct 31 '17 at 15:54

BENY

317,841
20
164
234

@nik, just do `boxplot(df)` – d.b Oct 31 '17 at 16:09
@nik edit and alos you can try the method d.b mentioned ~ :-) – BENY Oct 31 '17 at 16:09

score 1 · Answer 2 · edited Oct 31 '17 at 16:16

Using the melt() command would likely be the best solution assuming you've got an ID column as well. Check out this link for a further explanation.

newdf <- reshape2::melt(df, id="id").

If you don't have an ID already you can make rownames an ID column and then drop it later.

df$id <- rownames(df)
newdf <- melt(df, id="id").
newdf$id <- NULL

Also, it's bad form to ask further questions in the comments but run ?boxplot() in the console and it should be fairly straightforward.

How to reshape any data-frame to a 2-columns data-frame, with the (repeated) column names in the first and the corresponding values in the second?

2 Answers2