r dataframe and matrix result in different rownames when using rbind

Question

Method 1:

df1<-data.frame(A=1:5,B=2:6)
df2<-data.frame(A=1:5,B=2:6)
df3<-rbind(df1,df2)
row.names(df3)
 [1] "1"  "2"  "3"  "4"  "5"  "6"  "7"  "8"  "9"  "10"

Method 2:

df1<-data.frame(matrix(data = c(1:5,2:6),
                       nrow = 5,
                       dimnames = list(c(1:5),c("A","B"))))
df2<-data.frame(matrix(data = c(1:5,2:6),
                       nrow = 5,
                       dimnames = list(c(1:5),c("A","B"))))
df3<-rbind(df1,df2)
row.names(df3)
 [1] "1"  "2"  "3"  "4"  "5"  "11" "21" "31" "41" "51"

I understand that the rownames should be unique values, I would like to known why Mehtod1 and Method 2 generate different resutls in the rownames?

Answer 1

In Method 1, the row names are integers

df1 <- data.frame(A=1:5,B=2:6)
df2 <- data.frame(A=1:5,B=2:6)

dput(df1)
structure(list(A = 1:5, B = 2:6), class = "data.frame", row.names = c(NA,
-5L))

# or
attr(df1, "row.names")
[1] 1 2 3 4 5

and they stay integers after rbind

df3 <- rbind(df1, df2)

dput(df3)
structure(list(A = c(1L, 2L, 3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L),
    B = c(2L, 3L, 4L, 5L, 6L, 2L, 3L, 4L, 5L, 6L)), row.names = c(NA,
-10L), class = "data.frame")

In Method 2, the row names are strings

df1 <- data.frame(matrix(data = c(1:5,2:6),
                         nrow = 5,
                         dimnames = list(c(1:5),c("A","B"))))
df2 <- data.frame(matrix(data = c(1:5,2:6),
                         nrow = 5,
                         dimnames = list(c(1:5),c("A","B"))))

dput(df1)
structure(list(A = 1:5, B = 2:6), class = "data.frame", row.names = c("1",
"2", "3", "4", "5"))

# or
attr(df1, "row.names")
[1] "1" "2" "3" "4" "5"

and they also remain the original class, strings (character), after rbind

df3 <- rbind(df1, df2)

dput(df3)
structure(list(A = c(1L, 2L, 3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L),
    B = c(2L, 3L, 4L, 5L, 6L, 2L, 3L, 4L, 5L, 6L)), row.names = c("1",
"2", "3", "4", "5", "11", "21", "31", "41", "51"), class = "data.frame")

Strings can't be enumerated so they are made unique by adding an index to the string.

You can equalize the two by adding row.names=NULL when creating the data frame from a matrix with given row names.

df1 <- data.frame(matrix(data = c(1:5,2:6),
                         nrow = 5,
                         dimnames = list(c(1:5),c("A","B"))), row.names=NULL)

dput(df1)
structure(list(A = 1:5, B = 2:6), class = "data.frame", row.names = c(NA,
-5L))