R中独特观察表

Question

我有两个因子变量 - 在data.frame中称为"数据" - 看起来像这样:

brand  Country
 "A"    "ITA"
 "A"    "ITA"
 "C"    "SPA"
 "B"    "POR"
 "C"    "SPA"
 "B"    "POR"
 "A"    "ITA"
 "D"    "ITA"
 "E"    "SPA"
 "D"    "ITA"

我想获得一个表格,列出唯一的编号brands通过country.以下示例应该是:

# of unique brands  Country
        2             "ITA"
        2             "SPA"
        1             "POR"

首先,我试过:

data$var <- with(data, ave(brand, Country, FUN = function(x){length(unique(x))}))

但它不适用于因素,所以我转换了我的因素:

data$brand_t<-as.character(data$brand)
data$Country_t<-as.character(data$Country)

然后再说:

data$var <- with(data, ave(brand_t, Country_t, FUN = function(x){length(unique(x))}))

现在,如果我申请unique(data$var)我得到"2", "2", "1"哪个是正确的,但我无法得到我想要的表.可能很傻,但我无法解决.

我也想知道是否有一种更聪明的方法可以使用因子代替它.

再次感谢.

Answer 1

这里有两种快速方法,使用data.tablev> = 1.9.5或dplyr

library(data.table)
setDT(df)[, uniqueN(brand), by = Country]

要么

library(dplyr)
df %>%
  group_by(Country) %>%
  summarise(n = n_distinct(brand))

或者用基础R

aggregate(brand ~ Country, df, function(x) length(unique(x)))

要么

tapply(df$brand, df$Country, function(x) length(unique(x)))

或者如果您喜欢基本R简单语法并且您的数据集不是太大,您可以将方法组合在一起

aggregate(brand ~ Country, df, uniqueN) 

要么

aggregate(brand ~ Country, df, n_distinct)