R:根据与另一个数据框中的值匹配的列名,从一个数据框中提取行

时间:2019-06-17 17:46:41

标签: r

我想知道如何基于数据帧A中包含来自另一数据的多列值的某些列名来提取数据帧(数据帧A)中的一列中的值框(数据框B)

更具体地说。我有两个数据框:

数据框A 包含先天缺陷的组合。每行是不同的组合,每列是该组合中包含的缺陷数。

# Combinations data frame 
combos <- data.frame("combo_no"=c(1:4),
                     "Defect_A" = c(1,1,1,1),
                     "Defect_B" = c(3,2,3,4),
                     "Defect_C" = c(4,4,NA,7),
                     "Defect_D" = c(5,5,NA,8),
                     "Defect_E" = c(6,6,NA,NA))

数据框B 包含个别案例。第一列具有唯一标识符(CASE_ID)。其余各栏为特定的出生缺陷数,“ 1”代表“出生缺陷”,“ 0”代表“不存在”。

# Cases data set 
set.seed(99)
CASE_ID = c(1001:1005)
case1 = sample(0:1, 10, replace=TRUE)  
case2 = sample(0:1, 10, replace=TRUE)  
case3 = sample(0:1, 10, replace=TRUE)  
case4 = sample(0:1, 10, replace=TRUE)  
case5 = sample(0:1, 10, replace=TRUE)  
def<-data.frame(rbind(case1, case2, case3, case4, case5))
colnames(def)<- c(1:10)
cases<-cbind(CASE_ID,def)

所需的输出:我想从数据帧A中获得具有来自数据帧B的先天缺陷组合的CASE_ID列表。我还想指定存在哪种组合。 理想情况下,输出应如下所示:

# Desired Output
output <- data.frame("CASE_ID" = c(1002,1003),
                     "combo_no" = c(3,1))

谢谢您的帮助。

1 个答案:

答案 0 :(得分:0)

这里是解决方案,很长一段时间,以便逐步对其进行评论:

### my random generated cases DF:
cases
      CASE_ID 1 2 3 4 5 6 7 8 9 10
case1    1001 1 0 1 1 1 1 1 0 0  0
case2    1002 1 1 0 1 1 1 0 0 0  0
case3    1003 0 0 1 1 1 0 0 1 0  0
case4    1004 1 0 0 1 0 0 1 1 1  1
case5    1005 1 0 1 1 0 1 0 0 1  0
### initialize vectors to store found results
found_combos <- vector(); found_patients <- vector();
### open loop on combos rows
for (i in 1:nrow(combos)) {
  ### open empty vector to fill with the numbers that compose the defect
  defect_numbers <- vector()
  ### open loop on column and take the numbers
  for (col in colnames(combos)[2:length(colnames(combos))]) {
    number <- combos[i, col]
    if ( !is.na(number) ) defect_numbers <- append(defect_numbers, number)
  }
  ### sort the vector to avoid mismatch based on order
  defect_numbers <- sort( defect_numbers )
  ### open loop on patients table
  for ( pz in 1:nrow(cases) ) {
    pz_numbers <- sort( which( cases[pz,] == 1 )-1 )
    ### first condition: same length
    if ( length(pz_numbers) == length(defect_numbers) ) {
      ### same condition: exacly same numbers
      if (all(pz_numbers == defect_numbers)) {
        ### append to found results vectors
        found_patients <- append( found_patients, cases[pz,1] )
        found_combos <- append( found_combos, i )
      }
    }
  }
}

output <- data.frame("CASE_ID" = found_patients,
                     "combo_no" = found_combos)

### result:
output
  CASE_ID combo_no
1    1002        2

根据您的评论进行了编辑:

只需将条件从等于更改为%in%:

### initialize vectors to store found results
found_combos <- vector(); found_patients <- vector();
for (i in 1:nrow(combos)) {
  ### open empty vector to fill with the numbers that compose the defect
  defect_numbers <- vector()
  ### open loop on column and take the numbers
  for (col in colnames(combos)[2:length(colnames(combos))]) {
    number <- combos[i, col]
    if ( !is.na(number) ) defect_numbers <- append(defect_numbers, number)
  }
  ### sort the vector to avoid mismatch based on order
  defect_numbers <- sort( defect_numbers )
  ### open loop on patients table
  for ( pz in 1:nrow(cases) ) {
    pz_numbers <- sort( which( cases[pz,] == 1 )-1 )
    ### only condition: all defect_numbers in combo_numbers vector
    if (all(defect_numbers %in% pz_numbers)) {
      ### append to found results vectors
      found_patients <- append( found_patients, cases[pz,1] )
      found_combos <- append( found_combos, i )
    }
  }
}

output <- data.frame("CASE_ID" = found_patients,
                     "combo_no" = found_combos)