0

我有一个看起来像这样的数据表：

    Cause of Death                Ethnicity                 Count
1: ACCIDENTS EXCEPT DRUG POISONING ASIAN & PACIFIC ISLANDER  1368
2: ACCIDENTS EXCEPT DRUG POISONING                 HISPANIC  3387
3: ACCIDENTS EXCEPT DRUG POISONING       NON-HISPANIC BLACK  3240
4: ACCIDENTS EXCEPT DRUG POISONING       NON-HISPANIC WHITE  6825
5:              ALZHEIMERS DISEASE ASIAN & PACIFIC ISLANDER   285
---

我想创建一个新列，它只是不同种族之间死于特定死因的人的百分比。像这样：

   Cause of Death                Ethnicity                  Count  PercentofDeath
1: ACCIDENTS EXCEPT DRUG POISONING ASIAN & PACIFIC ISLANDER  1368     0.09230769
2: ACCIDENTS EXCEPT DRUG POISONING                 HISPANIC  3387     0.22854251
3: ACCIDENTS EXCEPT DRUG POISONING       NON-HISPANIC BLACK  3240     0.21862348
4: ACCIDENTS EXCEPT DRUG POISONING       NON-HISPANIC WHITE  6825     0.46052632
5:              ALZHEIMERS DISEASE ASIAN & PACIFIC ISLANDER   285     0.04049446
---

这是我的代码，非常难看：

   library(data.table)
   #load library, change to data table
   COD.dt <- as.data.table(COD)


   #function for adding the percent column
   lala  <- function(x){ 

   #see if I have initialized data.table I'm going to append to


      if(exists("started")){
        p <- COD.dt[x ==`Cause of Death`]
        blah <- COD.dt[x ==`Cause of Death`]$Count/sum(COD.dt[x ==`Cause of Death`]$Count)
        p$PercentofDeath <- blah
        started <<- rbind(started,p)
      }

       #initialize data table
        else{
            l <- COD.dt[x ==`Cause of Death`]
            blah <- COD.dt[x ==`Cause of Death`]$Count/sum(COD.dt[x ==`Cause of Death`]$Count)
            l$PercentofDeath <- (blah)
            started <<- l  
        }

 #if finished return
if(x == unique(COD.dt$`Cause of Death`)[length(unique(COD.dt$`Cause of Death`))]){
  return(started)
 }
 }

 #run function
 h <- sapply(unique(COD.dt$`Cause of Death`), lala)
  #remove from environment
 rm(started)
 #h is actually ends up being a list, the last object happen to be the one I want so I take that one
 finalTable <- h$`VIRAL HEPATITIS`

所以，如你所见。这段代码很丑陋，而且不适应。我希望从一些指导中了解如何使它变得更好。也许使用 dpylr 或其他一些功能？

最好的

4

2 回答 2

1

纯数据表解决方案也很简单，但这里是 dplyr：

 library(dplyr)

COD.dt %>% group_by(`Cause of Death`) %>%
    mutate(PercentofDeath = Count / sum(Count))

你可以把它变成一个函数，但它是一个很小的基本操作，大多数人都不会打扰。

于 2015-09-19T01:40:02.127 回答
0

我刚刚发现了一种更好的方法：

 library(data.table)
 #load library, change to data table
 COD.dt <- as.data.table(COD)

 #make column of disease total counts
 COD.dt[,disease:=sum(Count), by = list(`Cause of Death`)] 

 #use that column to make percents
 COD.dt[,percent:=Count/disease, by = list(`Cause of Death`)]
于 2015-09-19T01:42:43.043 回答