I'm working with census data. What the Data-set looks like is:
Household-Id Member-Type Education Birth
1 Father 12 1955
1 Mother 16 1963
1 Child 16 1986
1 Child 12 1995
2 Father 12 1950
2 Mother 9 1955
2 Child 18 1982
2 Child 14 1985
2 Child 16 1975
3 Father 16 1962
3 Mother 14 1965
3 Child 16 1990
What I want it to look like is:
Household-Id Member-Type Education Birth Mother-Education Birth-Order
1 Father 12 1955
1 Mother 16 1963
1 Child 16 1986 16 1
1 Child 12 1995 16 2
2 Father 12 1950
2 Mother 9 1955
2 Child 18 1982 9 1
2 Child 14 1985 9 2
2 Child 16 1975 9 3
3 Father 16 1962
3 Mother 14 1965
3 Child 16 1990 14 1
As far as I know, R doesn't support loop operation as in languages like Java or C, And I don't really have any idea on how to do this!
答案 0 :(得分：4)
Where did you hear that R doesn't support loops?? It most certainly does - this particular case sounds best suited for data.table (where loops are used internally)
install.packages("data.table")
library(data.table)
dat = as.data.table(YourDataFrame)
dat[Member.Type == "Child", Birth_Order:=rank(Birth) ,by=Household.Id]
dat[, MotherEducation := Education[Member.Type=="Mother"] , by=Household.Id]
dat[Member.Type != "Child", MotherEducation := NA]
dat
# Household.Id Member.Type Education Birth MotherEducation Birth_Order
# 1: 1 Father 12 1955 NA NA
# 2: 1 Mother 16 1963 NA NA
# 3: 1 Child 16 1986 16 1
# 4: 1 Child 12 1995 16 2
# 5: 2 Father 12 1950 NA NA
# 6: 2 Mother 9 1955 NA NA
# 7: 2 Child 18 1982 9 2
# 8: 2 Child 14 1985 9 3
# 9: 2 Child 16 1975 9 1
# 10: 3 Father 16 1962 NA NA
# 11: 3 Mother 14 1965 NA NA
# 12: 3 Child 16 1990 14 1
答案 1 :(得分：1)
Here's a dplyr approach:
library(dplyr)
dat = dat %>% group_by(Household.Id, Member.Type) %>%
arrange(Birth) %>%
mutate(Birth_Order = 1:n(),
Birth_Order = ifelse(Member.Type=="Child", Birth_Order, NA_integer_)) %>%
group_by(Household.Id) %>%
mutate(Mother_Education = ifelse(Member.Type=="Child",
Education[Member.Type=="Mother"], NA))
Household.Id Member.Type Education Birth Birth_Order Mother_Education
1 1 Child 16 1986 1 16
2 1 Child 12 1995 2 16
3 1 Father 12 1955 NA NA
4 1 Mother 16 1963 NA NA
5 2 Child 16 1975 1 9
6 2 Child 18 1982 2 9
7 2 Child 14 1985 3 9
8 2 Father 12 1950 NA NA
9 2 Mother 9 1955 NA NA
10 3 Child 16 1990 1 14
11 3 Father 16 1962 NA NA
12 3 Mother 14 1965 NA NA