在 R 中对多列使用排序和排名

Question

我正在尝试按每个州的最低费率对我的医院名称进行排名。当多家医院的费率相同时，应使用医院名称并按字母顺序排序来打破平局。到目前为止，我已经设法按州内的比率对它进行排名，按医院名称对其进行排序，但我不知道如何打破联系并在不跳过数字的情况下对其进行排名

到目前为止，这是我使用以下代码得到的结果：

outcome_data <- read.csv("outcome-of-care-measures.csv", na.strings="Not Available" ,stringsAsFactors=FALSE) #Read csv file
myData = outcome_data[,c(2, 7, 11)] #Retrieve only Hosp name, state and heart attack rate
arr1<-myData[complete.cases(myData[,3]),]  ##Remove NAs
arr2 <- arr1[order(arr1[2], arr1[3], arr1[1]),] #sort by state, then rate and then hospital name
arr3<-transform(arr2, rank = ave(rate, State, FUN = function(x) rank(x, ties.method = "min"))) #Rank by rate within each state

我目前得到的输出是：

Hospital.Name                           State  rate  rank
SOUTH PENINSULA HOSPITAL                AK     10.8  1
YUKON KUSKOKWIM DELTA REG HOSPITAL      AK     11.2  2
MAT-SU REGIONAL MEDICAL CENTER          AK     11.4  3
PEACEHEALTH KETCHIKAN MEDICAL CENTER    AK     11.4  3
ALASKA NATIVE MEDICAL CENTER            AK     11.6  5
BARTLETT REGIONAL HOSPITAL              AK     11.6  5
CENTRAL PENINSULA GENERAL HOSPITAL      AK     11.6  5
PROVIDENCE ALASKA MEDICAL CENTER        AK     12.4  8
ALASKA REGIONAL HOSPITAL                AK     13.4  9
FAIRBANKS MEMORIAL HOSPITAL             AK     15.6  10
GEORGE H. LANIER MEMORIAL HOSPITAL      AL     8.8   1
EVERGREEN MEDICAL CENTER                AL     9.1   2
BAPTIST MEDICAL CENTER EAST             AL     9.6   3
LAWRENCE MEDICAL CENTER                 AL     9.9   4
ANDALUSIA REGIONAL HOSPITAL             AL     10.1  5
JACKSON HOSPITAL & CLINIC INC           AL     10.2  6
BIRMINGHAM VA MEDICAL CENTER            AL     10.4  7
FLORALA MEMORIAL HOSPITAL               AL     10.4  7
GROVE HILL MEMORIAL HOSPITAL            AL     10.4  7
SPRINGHILL MEDICAL CENTER               AL     10.4  7
WEDOWEE HOSPITAL                        AL     10.4  7
PARKWAY MEDICAL CENTER                  AL     10.5  12
ST VINCENT'S BIRMINGHAM                 AL     10.6  13
WIREGRASS MEDICAL CENTER                AL     10.6  13
GADSDEN REGIONAL MEDICAL CENTER         AL     10.7  15
HALE COUNTY HOSPITAL                    AL     10.7  15
MOBILE INFIRMARY                        AL     10.7  15

但我想得到的是

Hospital.Name                           State  rate  rank
SOUTH PENINSULA HOSPITAL                AK     10.8  1
YUKON KUSKOKWIM DELTA REG HOSPITAL      AK     11.2  2
MAT-SU REGIONAL MEDICAL CENTER          AK     11.4  3
PEACEHEALTH KETCHIKAN MEDICAL CENTER    AK     11.4  4
ALASKA NATIVE MEDICAL CENTER            AK     11.6  5
BARTLETT REGIONAL HOSPITAL              AK     11.6  6
CENTRAL PENINSULA GENERAL HOSPITAL      AK     11.6  7
PROVIDENCE ALASKA MEDICAL CENTER        AK     12.4  8
ALASKA REGIONAL HOSPITAL                AK     13.4  9
FAIRBANKS MEMORIAL HOSPITAL             AK     15.6  10
GEORGE H. LANIER MEMORIAL HOSPITAL      AL     8.8   1
EVERGREEN MEDICAL CENTER                AL     9.1   2
BAPTIST MEDICAL CENTER EAST             AL     9.6   3
LAWRENCE MEDICAL CENTER                 AL     9.9   4
ANDALUSIA REGIONAL HOSPITAL             AL     10.1  5
JACKSON HOSPITAL & CLINIC INC           AL     10.2  6
BIRMINGHAM VA MEDICAL CENTER            AL     10.4  7
FLORALA MEMORIAL HOSPITAL               AL     10.4  8
GROVE HILL MEMORIAL HOSPITAL            AL     10.4  9
SPRINGHILL MEDICAL CENTER               AL     10.4  10
WEDOWEE HOSPITAL                        AL     10.4  11
PARKWAY MEDICAL CENTER                  AL     10.5  12
ST VINCENT'S BIRMINGHAM                 AL     10.6  13
WIREGRASS MEDICAL CENTER                AL     10.6  14
GADSDEN REGIONAL MEDICAL CENTER         AL     10.7  15
HALE COUNTY HOSPITAL                    AL     10.7  16
MOBILE INFIRMARY                        AL     10.7  17

有什么想法吗？

Answer 1

在 order 步骤

之后，我们需要一个分组序列号

library(dplyr)
arr2 %>%
     group_by(State) %>%
     mutate(rank = row_number())

或者如果我们从 'arr1'

开始

arr1 %>%
   arrange(State, rate,  Hospital.Name) %>%
   group_by(State) %>%
   mutate(rank = row_number())

或使用 base R

中的 ave

with(arr2, ave(seq_along(State), State, FUN = seq_along))
#[1]  1  2  3  4  5  6  7  8  9 10  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17

Answer 2

使用 data.table 这相对简单：

library(data.table)

# Read only relevant columns from csv file using data.table::fread
outcome_data <- fread("outcome-of-care-measures.csv",
                      na.strings="Not Available" ,
                      select = c("Hospital.Name","State","rate"))

# Drop rows NA values using data.table::na.omit
outcome_data <- na.omit(outcome_data)

## Use data.table::setkey to sort/index by State, then rate, then hospital name
setkey(outcome_data,State,rate,Hospital.Name)

## Add a rank column by state, order within groups will be based key order above
## (the .N operator is the number of rows in each State group)
outcome_data[,rank := seq_len(.N),by = .(State)]

在 R 中对多列使用排序和排名

Using sort and rank in R on multiple columns

sorting

r

rank