How do I transform multiple values in a dataframe,

问题:

Data:

names real_name type
a         A      AA
aa        A      AA
a-        A      AA
b         B      BB
bbb       B      BB
...

Wanted result:

names         real_name type
{a, aa, a-}      A       AA
{b, bbb}         B       BB

I want to get rid of the redundancy and be able to traverse through names. I have almost managed to get to the wanted results with loops and tidyr. What is the elegant way to get this?

回答1:

df <- tibble(names = c("a", "aa", "a-", "b", "bbb"), 
       real_name = c(rep("A", 3), rep("B", 2)),
       type = c(rep("AA", 3), rep("BB", 2)))

You can try nest if you want a list-col

df %>% nest(names)
# A tibble: 2 x 3
  real_name  type             data
      <chr> <chr>           <list>
1         A    AA <tibble [3 x 1]>
2         B    BB <tibble [2 x 1]>

Or summarize if you really want a string with all the names in it:

df %>% group_by(real_name, type) %>% summarize(x = paste(names, collapse = ", "))
# A tibble: 2 x 3
# Groups:   real_name [?]
  real_name  type         x
      <chr> <chr>     <chr>
1         A    AA a, aa, a-
2         B    BB    b, bbb