In R: How do I create a dataframe name from a string plus a column name plus categorical variable?

Question

Apologies, this seems like an easy question but I can't find the answer.

I'm using key word groups to search strings for important phrases. My table (srchtbl) classifies words by category (general thing they refer to) and component (actions vs. descriptions)

My method requires that I drill down to vectors to extract word groups to search. I'm able to create vectors for each category name and each component.

However, I also want to make dataframes for each category that are named by the category.

my data:

  word    pattern category component  
  <chr>   <chr>   <chr>    <chr>      
1 pack    pack    pkg      action     
2 protect protect pkg      action     
3 well    well    pkg      description
4 clever  clever  pkg      description
5 care    care    pkg      description
6 safe    safe    pkg      description

These statements create the appropriate dataframe with the appropriate name:

catgroups <- unique(srchtbl$category)

assign(paste("df_",  catgroups[i], sep = ""), srchtbl %>%  filter(category == catgroups[i]) %>% group_by(component))

which is fine, but how do I refer to it without using the whole statement? if I use:

print(paste("df_",  catgroups[3], sep = ""))

[1] "df_pkg"

So it's like I can't reference it again without using the entire assign statement.

Is there another way to concatenate a dataframe name and make a simple assignment, like:

"string" + catgroups[i] <- srchtbl %>%  filter(category == catgroups[3]) %>% group_by(component))

Ultimately the code will be looped so that the key word table can expand to any number of categories and components, so I don't want to type individual dataframe names

Answer 1

Consider base R's by or split which creates a named list of data frames from one or more grouping(s) where you can reference the individual data frames with $ or [[ qualifier. No need to flood global environment with many similarly structured objects. Instead maintain one list object. You lose no functionality of a data frame if stored in a list .

df_list1 <- split(srchtbl, srchtbl$category)
df_list1$pkg

#      word pattern category   component
# 1    pack    pack      pkg      action
# 2 protect protect      pkg      action
# 3    well    well      pkg description
# 4  clever  clever      pkg description
# 5    care    care      pkg description
# 6    safe    safe      pkg description


dflist2 <- by(srchtbl, srchtbl$category, identity)
dflist2[['pkg']]

#      word pattern category   component
# 1    pack    pack      pkg      action
# 2 protect protect      pkg      action
# 3    well    well      pkg description
# 4  clever  clever      pkg description
# 5    care    care      pkg description
# 6    safe    safe      pkg description