How to add a named vector as a row to a data frame

问题:

How to add a named vector to a data frame, with the components of the vector reordered according to the column names of the data frame?

I need to build a data frame one row at a time. A named vector is obtained by some processing and it provides the values for the row to be inserted. Problem is the named vector doesn't have components in the same order as data frame columns. This makes rbind produce wrong result. Here is the very simplified sample code:

df = data.frame(id=1:2, va=11:12, vb=21:22, vc=31:32)
v1 = c(id=4, va=14, vb=25, vc=NA)
df = rbind(df, v1)

So far, so good as this produces correct result. Now the next vector processing leads to:

v2 = c(va=19, id=9, vc=34, vb=NA)
df = rbind(df, v2)

This produces incorrect result. The correct result should be

id va vb vc
1  1 11 21 31
2  2 12 22 32
3  4 14 25 NA
4  9 19 NA 34

回答1:

Make a data frame out of v2 prior to the rbind:

rbind(df, as.data.frame(t(v2)))
##   id va vb vc
## 1  1 11 21 31
## 2  2 12 22 32
## 3  4 14 25 NA
## 4  9 19 NA 34

Here is why this works:

v2 has names, but it acts like a column vector to as.data.frame:

as.data.frame(v2)
##    v2
## va 19
## id  9
## vc 34
## vb NA

Thus, you must transpose the data to put it into the correct form:

as.data.frame(t(v2))
##   va id vc vb
## 1 19  9 34 NA

回答2:

You could reorder the vector

rbind(df, v2[names(df)])
  id va vb vc
1  1 11 21 31
2  2 12 22 32
3  9 19 NA 34


library(microbenchmark)
microbenchmark(rbind(df, v2[names(df)]),
               rbind(df, as.data.frame(t(v2))), times = 10000)
Unit: microseconds
                            expr     min      lq  median      uq      max neval
        rbind(df, v2[names(df)]) 212.773 219.305 222.572 294.895 15300.96 10000
 rbind(df, as.data.frame(t(v2))) 374.219 382.618 387.750 516.067 39951.31 10000