R中的條件交點

Question

尋找過濾掉兩個表中相交的動物（相交條件1）並在相同類別中跨表格共享相同大小（相交條件2）。知道一種有效的方法來編碼 - 例如，用dplyr？

library(dplyr) 
animal1 <- data.frame(type = c("cat", "dog", "dog","bird", "elephant"), 
         size = c("small","large","small", "medium", "large"), tableName = rep("animal1",5), stringsAsFactors = F) 
     #  type size tableName 
     # 1  cat small animal1 
     # 2  dog large animal1 
     # 3  dog small animal1 
     # 4  bird medium animal1 
     # 5 elephant large animal1 

animal2 <- data.frame(type = c("elephant", "dog", "dog", "elephant", "elephant"), 
         size = c("medium","large","large", "small", "large"), 
         tableName = rep("animal2",5), stringsAsFactors = F) 
     #  type size tableName 
     # 1 elephant medium animal2 
     # 2  dog large animal2 
     # 3  dog large animal2 
     # 4 elephant small animal2 
     # 5 elephant large animal2 


rbindAnimal <- rbind(animal1, animal2) 
     #  type size tableName 
     # 1  cat small animal1 
     # 2  dog large animal1 
     # 3  dog small animal1 
     # 4  bird medium animal1 
     # 5 elephant large animal1 
     # 6 elephant medium animal2 
     # 7  dog large animal2 
     # 8  dog large animal2 
     # 9 elephant small animal2 
     # 10 elephant large animal2 

# Intersection across both tables 
intersectType <- intersect(rbindAnimal %>% filter(tableName == "animal1") %>% select(type), 
              rbindAnimal %>% filter(tableName == "animal2") %>% select(type)) 
     #  type 
     # 1 elephant 
     # 2  dog 

rbindAnimal <- rbindAnimal[which(rbindAnimal$type %in% intersectType$type),] 

     #  type size tableName 
     # 2  dog large animal1 
     # 3  dog small animal1 
     # 5 elephant large animal1 
     # 6 elephant medium animal2 
     # 7  dog large animal2 
     # 8  dog large animal2 
     # 9 elephant small animal2 
     # 10 elephant large animal2 

# Needs to return row numbers! Here: 2,5,7,8, and 10 
#  type size tableName 
# 2  dog large animal1 
# 5 elephant large animal1 
# 7  dog large animal2 
# 8  dog large animal2 
# 10 elephant large animal2 

Answer 1

「需要返回行號！」

這是非常簡單的使用.I從data.table，存儲行號：

library(data.table) 
setDT(rbindAnimal) 

w <- rbindAnimal[, if (uniqueN(tableName) > 1L) .I, by=.(type, size)]$V1 
# [1] 2 7 8 5 10 
rbindAnimal[-w] 
#  type size tableName 
# 1:  cat small animal1 
# 2:  dog small animal1 
# 3:  bird medium animal1 
# 4: elephant medium animal2 
# 5: elephant small animal2 

相反的反連接（如OP的答案），我們只是用數字不包括行。

它是如何工作

• uniqueN計數唯一值的數量。 OP的條件是（釋義）：

  這兩種表格中都顯示了字體大小組合。

  其轉換爲

  uniqueN(tableName) > 1L的by=.(type, size)組行英寸

• if (cond) x給出x如果條件成立;否則，刪除組。

---

dplyr變種

它在dplyr正常工作，以及（雖然我不知道怎麼去行號）：

rbindAnimal %>% group_by(type, size) %>% filter(n_distinct(tableName) == 1L) 
#  type size tableName 
#  <chr> <chr>  <chr> 
# 1  cat small animal1 
# 2  dog small animal1 
# 3  bird medium animal1 
# 4 elephant medium animal2 
# 5 elephant small animal2 

Answer 2

解決方案：（！感謝合併尖@Imo）使用合併/ semi_join/anti_join

library(dplyr) 
animal1 <- data.frame(type = c("cat", "dog", "dog","bird", "elephant"), 
         size = c("small","large","small", "medium", "large"), tableName = rep("animal1",5), stringsAsFactors = F) 
     #  type size tableName 
     # 1  cat small animal1 
     # 2  dog large animal1 
     # 3  dog small animal1 
     # 4  bird medium animal1 
     # 5 elephant large animal1 

animal2 <- data.frame(type = c("elephant", "dog", "dog", "elephant", "elephant"), 
         size = c("medium","large","large", "small", "large"), 
         tableName = rep("animal2",5), stringsAsFactors = F) 
     #  type size tableName 
     # 1 elephant medium animal2 
     # 2  dog large animal2 
     # 3  dog large animal2 
     # 4 elephant small animal2 
     # 5 elephant large animal2 

rbindAnimal <- rbind(animal1, animal2) 
mergedAnimals <- merge(animal1, animal2, by = c("type","size"), all = T) 
sharedTypeSize <- mergedAnimals[complete.cases(mergedAnimals),] %>% select(type,size) %>% unique 
sharedTypeSize <- merge(rbindAnimal, sharedTypeSize) 

semi_join(rbindAnimal, sharedTypeSize) 
     #  type size tableName 
     # 1  dog large animal1 
     # 2  dog large animal2 
     # 3  dog large animal2 
     # 4 elephant large animal1 
     # 5 elephant large animal2 

anti_join(rbindAnimal, sharedTypeSize) 

     #  type size tableName 
     # 1  cat small animal1 
     # 2  dog small animal1 
     # 3  bird medium animal1 
     # 4 elephant medium animal2 
     # 5 elephant small animal2