CODEWITHSUNDEEP
functionfor-loopsapply
Filippo Marolla
Filippo Marolla

Reputation: 95

Loop over columns in a dataset in R

I have a dataset that looks like the one I create with the following code:

cluster <- rep(c(1,1,1,2,2,1,3,3,2,3,3))
measure_t1 <- rep(c(4.3, 4.7, 4.5, 3.4, 3.3, 4.7, 2.2, 2.1, 3.6, 2.2, 2.2))
measure_t2 <- rep(c(4.0, 4.1, 4.2, 3.5, 3.5, 4.6, 2.1, 2.4, 3.7, 2.3, 2.2))
measure_t3 <- rep(c(4.3, 4.2, 4.9, 3.8, 3.5, 4.3, 2.2, 2.8, 3.8, 2.7, 2.4))
df <- data.frame(cluster=cluster, measure_t1=measure_t1, measure_t2=measure_t2, 
measure_t3=measure_t3)

So I have measures of the same variable taken at three different sample occasions, in three different locations ("cluster"), with replicates within that location.

I basically want to add three columns to the dataset, where each new column contains the mean of each cluster for a given sample occasion. In other words, the new column "mean_t1" should contain the mean of measure_t_1 for cluster 1 every time cluster==1, the mean of measure_t1 for cluster 2 every time cluster==2, and the mean of measure_t1 for cluster 3 every time cluster==3. I want to do the same for measure_t2 and measure_t3.

I have no problems at doing it with the following code:

mean_t1 <- sapply(df$cluster, function(x) if(x==1) mean(df$measure_t1[df$cluster==1])
   else if(x==2) mean(df$measure_t2[df$cluster==2])
   else if(x==3) mean(df$measure_t2[df$cluster==3]))

and of course I can use the same code to create also mean_t2 and mean_t3, but my real dataset has many columns and this becomes time-consuming.

Therefore, I would like to do it iteratively, either in a loop or by making a function to be applied to the different columns, but I did not manage.

Upvotes: 1

Views: 39

Answers (1)

jdobres
jdobres

Reputation: 11957

We can accomplish this with minimal repetition using dplyr and the general tidyverse suite. Here we group by each cluster and compute the mean of each non-group column. Then rename the columns, and join to the original data set.

library(tidyverse)

df.means <- df %>% 
  group_by(cluster) %>% 
  mutate_all(mean)

colnames(df.means) <- gsub('measure', 'mean', colnames(df.means))

df.final <- bind_cols(df, df.means)

   cluster measure_t1 measure_t2 measure_t3 cluster1  mean_t1  mean_t2 mean_t3
1        1        4.3        4.0        4.3        1 4.550000 4.225000   4.425
2        1        4.7        4.1        4.2        1 4.550000 4.225000   4.425
3        1        4.5        4.2        4.9        1 4.550000 4.225000   4.425
4        2        3.4        3.5        3.8        2 3.433333 3.566667   3.700
5        2        3.3        3.5        3.5        2 3.433333 3.566667   3.700
6        1        4.7        4.6        4.3        1 4.550000 4.225000   4.425
7        3        2.2        2.1        2.2        3 2.175000 2.250000   2.525
8        3        2.1        2.4        2.8        3 2.175000 2.250000   2.525
9        2        3.6        3.7        3.8        2 3.433333 3.566667   3.700
10       3        2.2        2.3        2.7        3 2.175000 2.250000   2.525
11       3        2.2        2.2        2.4        3 2.175000 2.250000   2.525

Upvotes: 1

Related Questions