0

I am creating a function which will code people based on their family composition. As seen in the code below, #21 indicates a couple with with one or more children aged below 15. 

# 10    One adult without children
# 11    One adult with at least an own son or daughter aged less than 15
# 12    else: an own child aged 15 to 24 (1)
# 13    else: another child aged less than 15
# 20    One couple without children
# 21    One couple with at least:an own son or daughter aged less than 15
# 22    else: an own child aged 15 to 24 (1)
# 23    else: another child aged less than 15

Using the case_when() and between() functions I label observations as shown below: 

V(x)$age.code <- case_when(V(x)$age <= 15 ~ "x1",                      
                     between(V(x)$age, 15, 18) ~ "x2",
                     between(V(x)$age, 18, 24) ~ "X3", 
                     V(x)$age > 24 ~ "X4")

I then create the function to count adults/children by household. 

hhcomp <- function(x){
 if (sum(x == "X4", na.rm = T) == 1 | sum(x == "x3", na.rm = T) == 1)
 {y <- "10"}
 else if (sum(x == "X4", na.rm = T) | sum(x == "X3", na.rm = T) == 1 &  
                 sum(x  == "X1", na.rm = T) >= 1)
 {y <- "11"}
  else if ((sum(x == "X4", na.rm = T) | sum(x == "X3", na.rm = T) == 1) & 
                 sum(x == "X2", na.rm = T) >= 1 & sum(x == "X1") >= 1)
 {y <- "12"}
 else {y <- NA}
}

As seen above: 

X1 = less than 15 
x2 = 15 to 18
X3 = 18 to 24 
x4 = greater than 24

I am having trouble coding the sums in the hhcomp function, especially for #12 and #13. I am unsure of where to put the ampersand (&) and or (|) operators and if brackets are required

EDIT: 

ID FamID Age.Code 
1   A1     X1
2   A1     X3
3   A2     X4
4   A2     x4

df.2 <- df %>% group_by(FamID) %>% mutate(code = hhcomp(Age.Code)) %>% 
          ungroup()

I want to count the X terms for each FamID and apply the code to each ID as a new variable using mutate.

I am having problems with the logical operators

williamg15
  • 77
  • 7
  • Are you sure you can identify all those cases from age alone? For instance, how do you distinguish a couple without children from a single adult with a 20 years old child? – Julius Vainora Jan 04 '19 at 15:25
  • They are already grouped a family identifier. I will then use dplyr to write something along the lines of **df.2 <- df %>% group_by(famID) %>% mutate(code = hhcomp(age.code)) %>% ungroup()** – williamg15 Jan 04 '19 at 15:45
  • Absence of example data and such family identifiers complicate the situation, but possibly unnecessarily, since the question *seems* to be just about the logical operations. The meaning of X1, X2, X3, X3 is clear and it's also clear that `x` is a character vector of such values. So, perhaps the ultimate question could be formulated more specifically, in words, and with a minimal example and desired output? (Without requiring to know what "else" means and what extra variables there are.) – Julius Vainora Jan 04 '19 at 15:52
  • Hi i added an example and hopefully explained the question a little bit better – williamg15 Jan 04 '19 at 16:10
  • Can you add a reproducible example? It will make it easier for us to answer the question: https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example – william3031 Jan 04 '19 at 22:27

0 Answers0