How to loop through each column header in R and apply them to create graph

Question

My table is like this

family_group
col1 col2 col3
1    1    1 
1    2    NA
2    3    NA

I am trying to loop through each column header and call it in the function, so I can create a graph for each header

I tried this:

for (val in colnames(family_group))
{
  print(val)
  intermediate_tables <- count(na.omit(family_group$val))
  intermediate_frequency <-count(intermediate_tables$freq)
  plot  <- ggplot(data =intermediate_frequency, aes(x=intermediate_frequency$x, width = 0.5)) +geom_bar(aes(y=intermediate_frequency$freq), stat="identity", fill='lightgreen', color = 'green') + 
    geom_smooth(aes(y=intermediate_frequency$freq), method = 'loess') + 
    labs(x = "# of Samples in family", y = "# of families")
  plot
}

when printing val, it does show "col1", "col2", and "col3"

but then when I am trying to call family_group$val, it gives None

How do I call each column through looping?

Thanks

for your error with `family_group$val`, you should write `family_group[,val]`. BTW, I think the way you're plotting is also wrong but try to fix first your issue with `val`. — dc37, Nov 18 '19 at 21:27
yes that works, but like you said, the graph is not showing now, I was expecting 3 graph to pop up with the for loop — Norman Kuo, Nov 18 '19 at 21:30
On the last line of your `for` loop, try to replace `plot` by `print(plot)` — dc37, Nov 18 '19 at 21:37
[Don't use `$` inside `aes`](https://stackoverflow.com/questions/32543340/issue-when-passing-variable-with-dollar-sign-notation-to-aes-in-combinatio) — camille, Nov 18 '19 at 22:24
I don't know, but I also don't know that it's a bug to be fixed so much as an anti-pattern — camille, Nov 18 '19 at 22:41
I see, thanks in advance, I will define all my variable just in case for the future — Norman Kuo, Nov 18 '19 at 22:42

score 1 · Answer 1 · answered Nov 18 '19 at 21:43

An alternative to a for loop that you might find simpler, using the tidyverse function pivot_longer():

library(tidyverse)

family_group %>%
  pivot_longer(c(col1, col2, col3), 
               names_to = "column_name", 
               values_to = "value") %>%
  ggplot(aes(value)) +
  geom_bar() +
  facet_wrap(~ column_name)

The idea is rather than having multiple columns, all the data is reshaped into a very long but narrow dataframe. See more on SO here: How to plot all the columns of a data frame in R.

How to loop through each column header in R and apply them to create graph

1 Answers1