R - How to split/combine columns for multiple variables [duplicate] - r

Question

This question already has answers here:
R - reshaping 2 column data frame to multiple column matrix [duplicate]
(4 answers)
Reshaping 2 column data.table from long to wide
(2 answers)
Closed 5 years ago.
I'm quite the novice at R and I haven't been able to find an answer of how to split a column with multiple variables (sample 1-4) into separate columns whilst moving the data it correlates with. Here's an example:
Samples Content
Sample 1 70.7
Sample 1 91.6
Sample 1 92.6
Sample 1 65.2
Sample 1 80.0
Sample 1 82.1
Sample 1 88.1
Sample 1 92.2
Sample 1 53.3
Sample 1 80.0
Sample 1 60.3
Sample 1 89.7
Sample 1 84.8
Sample 1 94.0
Sample 1 71.8
Sample 1 76.9
Sample 1 91.4
Sample 1 57.9
Sample 1 61.9
Sample 1 71.5
Sample 2 88.7
Sample 2 67.6
Sample 2 61.7
Sample 2 70.8
Sample 2 45.3
Sample 2 55.6
Sample 2 64.6
Sample 2 62.7
Sample 2 72.4
Sample 2 46.8
Sample 2 59.0
Sample 2 63.7
Sample 2 67.0
Sample 2 71.6
Sample 2 48.3
Sample 2 55.6
Sample 2 62.5
Sample 2 60.0
Sample 2 72.9
Sample 2 47.4
Sample 3 42.3
Sample 3 48.2
Sample 3 64.0
Sample 3 33.3
Sample 3 19.0
Sample 3 41.0
Sample 3 53.1
Sample 3 46.5
Sample 3 30.0
Sample 3 43.4
Sample 3 43.7
Sample 3 92.0
Sample 3 53.0
Sample 3 33.0
Sample 3 48.4
Sample 3 43.2
Sample 3 41.8
Sample 3 62.5
Sample 3 33.3
Sample 3 49.3
Sample 4 51.8
Sample 4 57.3
Sample 4 43.3
Sample 4 42.3
Sample 4 37.6
Sample 4 54.9
Sample 4 71.1
Sample 4 33.8
Sample 4 43.1
Sample 4 39.1
Sample 4 63.0
Sample 4 74.0
Sample 4 31.0
Sample 4 48.3
Sample 4 42.9
Sample 4 62.2
Sample 4 35.4
Sample 4 33.8
Sample 4 40.7
Sample 4 41.2
I tried tidyr with no success. I want the output to be something like this;
Sample 1 Sample 2 Sample 3 Sample 4
70.7 88.7 42.3 51.8
91.6 67.6 48.2 57.3
92.6 61.7 64.0 43.3
65.2 70.8 33.3 42.3
80.0 45.3 19.0 37.6
82.1 55.6 41.0 54.9
88.1 64.6 53.1 71.1
92.2 62.7 46.5 33.8
53.3 72.4 30.0 43.1
80.0 46.8 43.4 39.1
60.3 59.0 43.7 63.0
89.7 63.7 92.0 74.0
84.8 67.0 53.0 31.0
94.0 71.6 33.0 48.3
71.8 48.3 48.4 42.9
76.9 55.6 43.2 62.2
91.4 62.5 41.8 35.4
57.9 60.0 62.5 33.8
61.9 72.9 33.3 40.7
71.5 47.4 49.3 41.2
Many thanks, if a solution is identified, is there an answer if I wanted to do the reciprocate?
Extra - Is there any way to preform a t-test on data which is stacked in one column such as the first example without having to transform it?

Answer 1

You may be having the "duplicate identifiers" issue using tidyr::spread. You need first to generate unique combinations of Sample + identifier, which you can do like this (assuming data frame named df1):
library(tidyverse) # for dplyr + tidyr
df1 %>%
group_by(Samples) %>%
mutate(id = row_number()) %>%
spread(Samples, Content) %>%
select(-id)
"if I wanted to do the reciprocate"
Do you mean go the other way, from the wide form back to the original long form? Then you use gather. Add this to the end of the code above and see what happens:
%>% gather(Samples, Content)
t-test: there are lots of ways you could run a t-test on the long format data. For example, a base R way to compare Samples 1 and 2 might be:
t.test(df1[df1$Samples == "Sample 1", "Content"],
df1[df1$Samples == "Sample 2", "Content"])

Answer 2

As the number of elements for each 'Sample' is the same, we can use unstack from base R
unstack(df1, Content~Samples)
# Sample.1 Sample.2 Sample.3 Sample.4
#1 70.7 88.7 42.3 51.8
#2 91.6 67.6 48.2 57.3
#3 92.6 61.7 64.0 43.3
#4 65.2 70.8 33.3 42.3
#5 80.0 45.3 19.0 37.6
#6 82.1 55.6 41.0 54.9
#7 88.1 64.6 53.1 71.1
#8 92.2 62.7 46.5 33.8
#9 53.3 72.4 30.0 43.1
#10 80.0 46.8 43.4 39.1
#11 60.3 59.0 43.7 63.0
#12 89.7 63.7 92.0 74.0
#13 84.8 67.0 53.0 31.0
#14 94.0 71.6 33.0 48.3
#15 71.8 48.3 48.4 42.9
#16 76.9 55.6 43.2 62.2
#17 91.4 62.5 41.8 35.4
#18 57.9 60.0 62.5 33.8
#19 61.9 72.9 33.3 40.7
#20 71.5 47.4 49.3 41.2
No external packages are used
If the number of 'Sample' elements are different, then dcast from data.table can be used (works in both cases)
library(data.table)
dcast(setDT(df1), rowid(Samples)~Samples, value.var = "Content")