如何使用另一个R数据帧的值对一个R数据帧进行子集化？

Question

I have two dataframes in R:

Died.At <- c(22,40,72,41, ...)
Writer.At <- c(16, 18, 36, 36)
Name <- c("John Doe", "Edgar Poe", "Walt Whitman", "Jane Austen", ...)
Gender <- c("MALE", "MALE", "MALE", "FEMALE", ...)
Date.Of.Death <- c("2015-05-10", "1849-10-07", "1892-03-26","1817-07-18", ...)
Pet <- c("cat", "dog", "cat", "cat")
df1 = data.frame(Died.At, Writer.At, Name, Gender, Pet)
print(df1)
  Died.At Writer.At     Name          Gender    Pet
1      22        16     John Doe      MALE      cat
2      40        18     Edgar Poe     MALE      dog
3      72        36     Walt Whitman  MALE      cat
4      41        36     Jane Austen   FEMALE    cat
.....

In df1 not each row for Name is unique (ie there are several rows with the same author.)

The second dataframe df2 , there is also a column Name with both authors from df1 (eg Jane Austen) and completely new authors. This dataframe is also far larger.

print(length(unique(df1$Name)))
## output 1168
print(length(unique(df2$Name)))
## output 5572

I would like to subset df2 such that the only names are the names from df1 .

My idea was to do this:

subset_df2 = df2[df2$Name == unique(df1$Name)]

However, I would expect there to be 1168 unique author names here:

print(length(unique(subset_df2$Name)))
## output 880

That's less than I was expecting. Where is my error?

Answer 1

You can use match(df2$Name, df1$Name) or df2$Name %in% df1$Name which return a vector of logicals the length of df2$Name, and logical TRUE where df2$Name is in df1$Name. You can then use this to index df2.

subset_df2 <- df2[df2$Name %in% df1$Name, ]

See ?match

As for why your code did not work, please see the output of this exercise:

a = LETTERS[sample(1:10, size=15, replace=T)]
b = c(unique(a), LETTERS[15:30])
# compare
b == unique(a) 
b[b == unique(a)]
# vs
b %in% a
b[b %in% a]

Note also b %in% a is not equivalent to a %in% b and therefore b[a %in% b] would yield an incorrect result.

Furthermore, when indexing a data frame you need to provide a row range and column range.