How to label values in a column based on a dataframe "dictionary"

Question

I have data like this

and a dataframe like this

col1   col2
12     "London"
33     "Paris"
924    "Singapore"

How do I label the first dataframe based on the columns in the second dataframe using r's labelled?

I know that using val_labels() I can apply a value label to values in a column using:

val_labels(df$col1) <- c(London = 12, Paris = 33, Singapore = 924)

But I have 1000 different values and need an approach that allows me to use a dataframe to do it.

`labelled` package is used to label variables (columns), not rows — Waldi, Mar 10 '22 at 13:01
What do you want? Just get a col2 in your first data.frame with the city names? — Merijn van Tilborg, Mar 10 '22 at 13:18
@MerijnvanTilborg I have to export it after to spss that works with value labels and it is helpful to use labels for data analysis, because the data analysis tools don't work with strings but can use labels to assign names in the figures so I don't have to do it manually. — Victor Nielsen, Mar 10 '22 at 13:58

score 0 · Accepted Answer · answered Mar 10 '22 at 14:07

You can try this, though there are likely more elegant solutions:

df <- data.frame(col1 = c(33, 924, 33, 12, 924))
dic <- data.frame(col1 = c(12, 33, 924),
                  col2 = c("London","Paris","Singapore"))

library(labelled)
ct <- 1
for(i in dic$col1){
  val_label(df, i) <- as.character(dic[ct,2])
  ct <- ct+1
}
str(df)
# > str(df)
# 'data.frame': 5 obs. of  1 variable:
#   $ col1: dbl+lbl [1:5]  33, 924,  33,  12, 924
# ..@ labels: Named num  12 33 924
# .. ..- attr(*, "names")= chr [1:3] "London" "Paris" "Singapore"
# >

score 0 · Answer 2 · answered Mar 10 '22 at 14:18

0

With val_labels:

library(labelled)

col1=c(33,924,33,12,924)

ref <- read.table(text='
col1   col2
12     "London"
33     "Paris"
924    "Singapore"',header=T)

val_labels(col1)<-setNames(ref$col1,ref$col2)

col1
#> <labelled<double>[5]>
#> [1]  33 924  33  12 924
#> 
#> Labels:
#>  value     label
#>     12    London
#>     33     Paris
#>    924 Singapore

answered Mar 10 '22 at 14:18

Waldi

39,242
6
30
78

I get this error: Error in `stop_vctrs()`: ! Can't convert `labels` to match type of `x` . – Victor Nielsen Mar 10 '22 at 16:59
do you get the error with the above dataset, or on other data? This is working fine for me on `R 4.1.0` with `reprex` package – Waldi Mar 10 '22 at 18:37
what is `class(col1)` for you, `integer` as in my code or `character`? both `col1` should have the same datatype – Waldi Mar 10 '22 at 18:43
They are both integer. I couldn't get it to work, but your solution works for the data I provided, so I can't ask for more. – Victor Nielsen Mar 11 '22 at 11:37

How to label values in a column based on a dataframe "dictionary"

2 Answers2