如何根据字典更改或替换行名

Question

我想根据提供的ICD-10代码字典，将一行中的ICD-10代码更改为疾病的名称。

这是初始数据

id <- c("1","2","3")
Dx1 <- c("E119", "I251","I20")
Dx2 <- c("I20", "I251","E119")
Dx3 <- c("I251", "E119","I20")
df <- data.frame(id,Dx1,Dx2,Dx3)
df

这是ICD-10代码字典，在本例中有3个代码，但实际上，ICD-10代码包含9万4千个代码。

ICD <- c("I251", "E119","I20")
Disease <- c("Acute Myocard Infarct", "Type 2 Diabetes", "Chest Pain")
CodeDictionary <- data.frame(ICD,Disease)
CodeDictionary

这是我的目标

id <- c("1","2","3")
Dx1 <- c("Type 2 Diabetes", "Acute Myocard Infarct","Chest Pain")
Dx2 <- c("Chest Pain", "Acute Myocard Infarct","Type 2 Diabetes")
Dx3 <- c("Acute Myocard Infarct", "Type 2 Diabetes","I20")
dfGoal <- data.frame(id,Dx1,Dx2,Dx3)
dfGoal

我尝试了dplyr中的内连接，但它不起作用。谢谢你的帮助!

Answer 1

您可以使用plyr包通过函数mapvalues(...)来完成此操作。但是，您必须遍历列，这并不理想。

library(plyr)

id <- c("1","2","3")
Dx1 <- c("E119", "I251","I20")
Dx2 <- c("I20", "I251","E119")
Dx3 <- c("I251", "E119","I20")
df <- data.frame(id,Dx1,Dx2,Dx3)

dfGoal <- df

for(i in c(2:dim(df)[2])){
  dfGoal[,i] <- mapvalues(df[,i],
  from = c("I251", "E119","I20"),
  to = c("Acute Myocard Infarct", "Type 2 Diabetes", "Chest Pain"))
}

dfGoal

在这个线程中，您可能会发现更多有用的技术：https://stackoverflow.com/a/25790005/7120715

Answer 2

我认为inner_join是正确的方向。在此之前，您需要使用pivot_longer，然后一旦您有了疾病名称，您就可以使用pivot_wider来获取您的dfGoal：

library(tidyverse)

id <- c("1","2","3")
Dx1 <- c("E119", "I251","I20")
Dx2 <- c("I20", "I251","E119")
Dx3 <- c("I251", "E119","I20")
df <- data.frame(id,Dx1,Dx2,Dx3)
df


ICD <- c("I251", "E119","I20")
Disease <- c("Acute Myocard Infarct", "Type 2 Diabetes", "Chest Pain")
CodeDictionary <- data.frame(ICD,Disease)
CodeDictionary

id <- c("1","2","3")
Dx1 <- c("Type 2 Diabetes", "Acute Myocard Infarct","Chest Pain")
Dx2 <- c("Chest Pain", "Acute Myocard Infarct","Type 2 Diabetes")
Dx3 <- c("Acute Myocard Infarct", "Type 2 Diabetes","I20")
dfGoal <- data.frame(id,Dx1,Dx2,Dx3)
dfGoal

df_goal <- df %>%
  pivot_longer(-id, names_to = "diagnoses", values_to = "ICD") %>%
  inner_join(CodeDictionary, by = "ICD") %>%
  select(id, diagnoses, Disease) %>%
  pivot_wider(names_from = diagnoses, values_from = Disease)

希望这能有所帮助！