andkov
commented
8 years ago Objective
We need to de construct the names of the melted variables into individual components, which would create values for the new variables. Consider the data:
variable
1 intage2000
2 intage2005
3 intage2009
4 dressing.70
5 dressing.75
6 dressing.79
7 feed.70
8 feed.75
9 feed.79
10 dem2000
11 dem2005
12 dem2009
13 AE69
14 AG58
15 AH108
16 AE320
17 AG384
18 AH414Solution
ds %>% dplyr::arrange(variable)
regex <- "^(\\w+?)\\.(\\d{2})$"
regex2 <- "^(AE|AG|AH)(\\d+?)$"
regex3 <- "^(\\w+?)(\\d{4})$"
d <- ds %>% dplyr::mutate(
age = variable, # copy existing
age = gsub(regex, "\\2", variable),
age = gsub(regex2, "\\1", age),
age = gsub(regex3, "\\2", age),
age = car::recode(age, " c('2000','AE')='70';
c('2005','AG')='75';
c('2009','AH')='79';"),
measure = gsub(regex, "\\1", variable),
measure = gsub(regex2, "\\2", measure),
measure = gsub(regex3, "\\1", measure),
measure = car::recode(measure, " c('320','384','414')='smoking';
c('69','58','108')='bmi' "),
year = age + 1930
) %>% dplyr::arrange(desc(variable))
(d)
variable age measure year
1 intage2009 79 intage 2009
2 intage2005 75 intage 2005
3 intage2000 70 intage 2000
4 feed.79 79 feed 2009
5 feed.75 75 feed 2005
6 feed.70 70 feed 2000
7 dressing.79 79 dressing 2009
8 dressing.75 75 dressing 2005
9 dressing.70 70 dressing 2000
10 dem2009 79 dem 2009
11 dem2005 75 dem 2005
12 dem2000 70 dem 2000
13 AH414 79 smoking 2009
14 AH108 79 bmi 2009
15 AG58 75 bmi 2005
16 AG384 75 smoking 2005
17 AE69 70 bmi 2000
18 AE320 70 smoking 2000
Agenda