с помощью ограничений определить уникальные идентификаторы

Question

Person <- c(1,2,3) 
Age <- c(10,22,30) 
Height <- c(140,185,160) 
Weight <- c(65, 80, 75) 

People <- data.frame(Person, Age, Height, Weight) 

Age_cats_type1 [5-15], [20-30], [35-45] 
Age_cats_type2 [8-13], [14-16], [18-40] 

Height_cat_Type1 [100-120], [121-140], [141-186] 
Height_cat_type2 [110-125], [126-145], [146-190] 

Weight_cat_Type1 [50-60], [61-78], [79-85] 
Weight_cat_Type2 [55-75], [76-90], [91-100] 

для людей [1,2] (возраст = 10), это вписывается в Age_cats_type1==1 и Age_cats_type2==1. для людей [1,3] (высота = 140), это вписывается в Height_cat_Type1==2 и Height_cat_Type2==2

Теперь я хочу, чтобы создать таблицу для каждого уникального возможного результата интервалов для (Age_cats_type1==1)|(Age_cats_type1==2), (Height_cats_type1==1)|(Height_cats_type1==2), (Weight_cats_type1==1)|(Weight_cats_type1==2).

Желаемый результат должен выглядеть как желтое изображение ниже. Приведенная выше таблица представляет собой краткое изложение возможностей для каждого invterval

Это тесно связано с following question, но если вы будете следовать коды, используемые здесь, как указано на ошибки BrodieG возникают на третьей итерации. В этом примере мы используем foverlaps в data.table

Я использовал следующий код

library(intervals) 
# create our limits 

AGE_cats_type1 <- Intervals(
    matrix(c(5, 15, 20, 30, 35, 40), ncol = 2, byrow = TRUE), 
    closed = c(TRUE, T), 
    type = "Z" 
) 

AGE_cats_type2 <- Intervals(
    matrix(c(8, 13, 14, 16, 18, 40), ncol = 2, byrow = TRUE), 
    closed = c(TRUE, T), 
    type = "Z" 
) 

Height_cats_type1 <- Intervals(
    matrix(c(100, 120, 121, 140, 141, 186), ncol = 2, byrow = TRUE), 
    closed = c(TRUE, T), 
    type = "Z" 
) 

Height_cats_type2 <- Intervals(
    matrix(c(110, 125, 126, 145, 146, 190), ncol = 2, byrow = TRUE), 
    closed = c(TRUE, T), 
    type = "Z" 
) 

Weight_cats_type1 <- Intervals(
    matrix(c(50, 60, 61, 78, 79, 85), ncol = 2, byrow = TRUE), 
    closed = c(TRUE, T), 
    type = "Z" 
) 

Weight_cats_type2 <- Intervals(
    matrix(c(55, 75, 76, 90, 91, 100), ncol = 2, byrow = TRUE), 
    closed = c(TRUE, T), 
    type = "Z" 
) 

#now format data 
# first for age 
library(data.table) 
PEOPLE1 <- data.table(People) 
PEOPLE1[, A1:=Age] 

I_age_1 <- data.table(cbind(data.frame(AGE_cats_type1), idX=1:3, idY=0)) 
I_age_2 <- data.table(cbind(data.frame(AGE_cats_type2), idX=0, idY=1:3)) 

setkey(I_age_1, X1, X2) 
setkey(I_age_2, X1, X2) 

PEOPLE2 <- data.frame(rbind(
    foverlaps(PEOPLE1, I_age_1, by.x=c("Age", "A1"), nomatch=0), 
    foverlaps(PEOPLE1, I_age_2, by.x=c("Age", "A1"), nomatch=0))) 

#################################################### 
# second iteration for height 

PEOPLE3 <- data.table(PEOPLE2) 
PEOPLE3[, B1:=Height] 

I_height_1 <- data.table(cbind(data.frame(Height_cats_type1), idXa=1:3, idYa=0)) 
I_height_2 <- data.table(cbind(data.frame(Height_cats_type2), idXa=0, idYa=1:3)) 

setkey(I_height_1, X1, X2) 
setkey(I_height_2, X1, X2) 

PEOPLE4 <- data.frame(rbind(
    foverlaps(PEOPLE3, I_height_1, by.x=c("Height", "B1"), nomatch=0), 
    foverlaps(PEOPLE3, I_height_1, by.x=c("Height", "B1"), nomatch=0))) 

################################################ 
# third iteration 

PEOPLE5 <- data.table(PEOPLE4) 
PEOPLE5[, C1:=Weight] 

I_weight_1 <- data.table(cbind(data.frame(Weight_cats_type1), idXb=1:3, idYb=0)) 
I_weight_2 <- data.table(cbind(data.frame(Weight_cats_type2), idXb=0, idYb=1:3)) 

setkey(I_weight_1, X1, X2) 
setkey(I_weight_2, X1, X2) 

PEOPLE6 <- data.frame(rbind(
    foverlaps(PEOPLE5, I_weight_1, by.x=c("Height", "B1"), nomatch=0), 
    foverlaps(PEOPLE5, I_weight_2, by.x=c("Height", "B1"), nomatch=0))) 

Но получить ошибку в человек6.

Error in setcolorder(ans, c(xcols1, ycols, xcols2)) : 
    neworder is length 16 but x has 18 columns. 

Когда я смотрю на PEOPLE4 мы видим IDX IdY idxA и idyA являются Age_cats_type1, Age_cats_type2, Height_cat_Type1 и Height_cat_Type2 значения

Answer 1

У вас есть несколько проблем с формулировкой Вашего вопроса. Давайте попробуем восстановить то, что вы имели в виду.

# Let's first construct the data 
Person <- c(1,2,3) 
Age <- c(10,22,30) 
Height <- c(140,185,160) 
Weight <- c(65, 80, 75) 

People <- data.frame(Person, Age, Height, Weight) 

# Results in something like this: 
# Person Age Height Weight 
# 1  1 10 140  65 
# 2  2 22 185  80 
# 3  3 30 160  75 

# Now we want to represent ranges. One way to do it would be: 

Age_cats_type1 <- list(c(5, 15), c(20, 30), c(35, 45)) 
Age_cats_type2 <- list(c(8, 13), c(14, 16), c(18, 40)) 

Height_cat_Type1 <- list(c(100, 120), c(121, 140), c(141, 186)) 
Height_cat_Type2 <- list(c(110, 125), c(126, 145), c(146, 190)) 

Weight_cat_Type1 <- list(c(50, 60), c(61, 78), c(79, 85)) 
Weight_cat_Type2 <- list(c(55, 75), c(76, 90), c(91, 100)) 

# Then you mentioned something like People[1,1] meant age == 10. 
# I believe you made a mistake here. If you type People[1, 1] in the console, 
# you'll find you that it returns Person == 1. Therefore, I think 
# that you meant to say the People data frame was constructed without the Person vector. 

People <- data.frame(Age, Height, Weight) 

# Now People[1, 1] returns age == 10. 

# Then you went on to say that you wanted some function that returned Age_cats_type == 1 
# Well, it seems that you want the first element of the list of ranges that contains the specified value. 
# Then let's build it 

contains_value <- function(range, value) { 
    lower <- range[1] 
    upper <- range[2] 
    lower <= value && value <= upper 
} 

range_index <- function(ranges_list, value){ 
    which(sapply(ranges_list, contains_value, value))[1] 
} 

range_index(Age_cats_type1, People[1, 1]) # 1 
range_index(Age_cats_type2, People[1, 1]) # 1 
range_index(Height_cat_Type1, People[1, 2]) # 2 
range_index(Height_cat_Type2, People[1, 2]) # 2 

# Now I didn't understand what the table you were trying to construct was, but maybe these functions will help you build it.