如何提取另一个数据帧的几个观测值并将其包含在条件下的主数据帧中？

Question

我正在使用两个不同的数据框，这两个数据框包括对特定年份的分类变量及其频率的不同观察，但是对于两个不同大小的不同组（第2组大于第1组）。

现在，我想包括第1组中第2组的观察频率，这些观察结果与分类变量（a）具有相同的“值”，并且属于同一年。到目前为止，我无法找到正确的函数或循环。

到目前为止我尝试过的：

library(tidyverse)
    library(dplyr)

    year_1 <- c(1914, 1915, 1916, 1917)
    a_1 <- c("blue", "green", "green", "blue")
    df_1 <- data.frame(a_1, year_1)
    df_1 <- group_by(a_1, year_1) %>% count(name = "counts") %>% 
            ungroup(a_1, year_1) %>% 
    df_1 <- mutate(freq_1=counts/sum(counts))

    year_2 <- c(1912, 1913, 1914, 1915, 1916, 1917, 1918)
    a_2    <- c("black", "pink", "blue", "green", "green", "pink", "blue")
    df_2 <- data.frame(a_2, year_2)
    df_2 <- group_by(a_2, year_2) %>% count(name = "counts") %>% 
            ungroup(a_2, year_2) %>% 
    df_2 <- mutate(freq_2=counts/sum(counts))



    for(i in df_1) {
      if(df_2$a_2==df_1$a_1[i] & df_2$year_2==df_1$year_1[i]) {
       df_1 %>% 
       mutate(df_1, freq_2 = df_2$freq_2) %>% 
        return(df_1)}

      }

Answer 1

我不知道这是不是你想要的：

library(dplyr)

#just recreating your data with minor corrections
year_1 <- c(1914, 1915, 1916, 1917)
a_1 <- c("blue", "green", "green", "blue")
df_1 <- data.frame(a_1, year_1, stringsAsFactors = FALSE)
df_1 <- df_1 %>% group_by(a_1, year_1) %>% tally() %>% 
        ungroup(a_1, year_1) %>% 
        mutate(freq_1=n/sum(n))

year_2 <- c(1912, 1913, 1914, 1915, 1916, 1917, 1918)
a_2    <- c("black", "pink", "blue", "green", "green", "pink", "blue")
df_2 <- data.frame(a_2, year_2, stringsAsFactors = FALSE)
df_2 <- df_2 %>% group_by(a_2, year_2) %>% tally() %>% 
        ungroup(a_2, year_2) %>% 
        mutate(freq_2=n/sum(n))


df_1 %>% left_join(df_2, by = c("year_1" = "year_2", "a_1" = "a_2")) %>%
  mutate(count = ifelse(!is.na(n.y), n.x + n.y, n.x),
         freq = count/sum(count)) %>% select(-n.y, -n.x, - freq_1, -freq_2 )

# A tibble: 4 x 4
  a_1   year_1 count  freq
  <chr>  <dbl> <int> <dbl>
1 blue    1914     2 0.286
2 blue    1917     1 0.143
3 green   1915     2 0.286
4 green   1916     2 0.286

这会将df_2和df_1中的所有类别添加到同一年的两个数据框中，从两个数据框中相加各自的频率，并计算一个类似于freq_1and和freq_2列的新百分比。

Answer 2

我会用aggregate和merge或ave这样做：

#Your original Data
df1 <- data.frame(a=c("blue", "green", "green", "blue"), year=c(1914, 1915, 1916, 1917))
df2 <- data.frame(a=c("black", "pink", "blue", "green", "green", "pink", "blue"), year=c(1912, 1913, 1914, 1915, 1916, 1917, 1918))

#Count per a and year
df1 <- with(df1, aggregate(list(count=year), list(a=a, year=year), FUN=length))
df2 <- with(df2, aggregate(list(count=year), list(a=a, year=year), FUN=length))

#Include counts of df2 in df1 using aggregate and merge
merge(df1[c("a","year")], aggregate(count ~ a + year, data=rbind(df1, df2), FUN=sum))
      a year count
1  blue 1914     2
2  blue 1917     1
3 green 1915     2
4 green 1916     2

#Include counts of df2 in df1 using ave
df1$count <- with(rbind(df1, df2), ave(count, a, year, FUN=sum))[1:NROW(df1)]
df1
      a year count
1  blue 1914     2
2 green 1915     2
3 green 1916     2
4  blue 1917     1