Basically, I had an initial data.frame where I used participants name as row.names. For various reasons, I had to stack then merge the data in the long format and then based on a factor variable I selected only a proportion of that data. And then I aggregated this so I have the wide data again with participants as row.names. However, the order is not the same. so if I wanted to cbind the original wide data with the new data it would mess up my data as the order of row.names is different. I tried sort(), order(), transform() and read several questions, but I couldn't find a way of doing it.
DF1
> V1 V2 V3
> AAA 24 22 37
> BBB 21 22 33
> CCC 30 32 38
> DDD 21 23 35
The other data frame has the same dim(), however different order of the row.names.
DF2 row.names
> BBB
> CCC
> AAA
> DDD
I would like to sort DF1 based of rownames(DF2) so that the variables stay with the rownames, so I can then cbind (D1, D2) and because the rownames are equal, the same variable would belong to the same participant. I might have overcomplicated it, sorry :)
Basically, I want to reorder the whole data.frame(DF1) based on rownames in DF2.
I know it's a lame question, but I couldn't find a working answer.
答案 0 :(得分：7)
Here's one option, using the built-in mtcars data frame for illustration:
# Create new sorted data frame. This is analogous to the second data frame
# by which you want to sort the first one.
mtcars.sorted = mtcars[order(rownames(mtcars)),]
# Sort original data frame by the order of the new data frame
mtcars[match(rownames(mtcars.sorted), rownames(mtcars)),]
So, in your case, the code would be:
DF1[match(rownames(DF2), rownames(DF1)), ]
答案 1 :(得分：3)
You can use match
> df <- data.frame(a=c("AAA","BBB","CCC"),b=1:3,c=6:8,d=9:11)
> df
a b c d
1 AAA 1 6 9
2 BBB 2 7 10
3 CCC 3 8 11
> df1 <- data.frame(a=c("CCC","AAA","BBB"))
> df1
a
1 CCC
2 AAA
3 BBB
> Final_df <- df[match(df1$a,df$a),] ## Here it is
> Final_df
a b c d
3 CCC 3 8 11
1 AAA 1 6 9
2 BBB 2 7 10