Merge or combine by rownames

Question

In the example below I have two datasets (Z and A). I want to merge or combine these sets by the ILMN numbers. If there is no match, fill in NA.

z <- matrix(c(0,0,1,1,0,0,1,1,0,0,0,0,1,0,1,1,0,1,1,1,1,0,0,0,"RND1","WDR", "PLAC8","TYBSA","GRA","TAF"), nrow=6,
    dimnames=list(c("ILMN_1651838","ILMN_1652371","ILMN_1652464","ILMN_1652952","ILMN_1653026","ILMN_1653103"),c("A","B","C","D","symbol")))

t<-matrix(c("GO:0002009", 8, 342, 1, 0.07, 0.679, 0, 0, 1, 0, 
        "GO:0030334", 6, 343, 1, 0.07, 0.065, 0, 0, 1, 0,
        "GO:0015674", 7, 350, 1, 0.07, 0.065, 1, 0, 0, 0), nrow=10, dimnames= list(c("GO.ID","LEVEL","Annotated","Significant","Expected","resultFisher","ILMN_1652464","ILMN_1651838","ILMN_1711311","ILMN_1653026")))

The result will be like this:

             [,1]         [,2]         [,3]         [,4]
GO.ID        "GO:0002009" "GO:0030334" "GO:0015674"  NA
LEVEL        "8"          "6"          "7"           NA
Annotated    "342"        "343"        "350"         NA
Significant  "1"          "1"          "1"           NA
Expected     "0.07"       "0.07"       "0.07"        NA
resultFisher "0.679"      "0.065"      "0.065"       NA
ILMN_1652464 "0"          "0"          "1"           PLAC8
ILMN_1651838 "0"          "0"          "0"           RND1
ILMN_1711311 "1"          "1"          "0"           NA
ILMN_1653026 "0"          "0"          "0"           GRA

Joris Meys · Answer 1 · 2014-09-08T11:32:11.960

71

Using merge and renaming your t vector as tt (see the PS of Andrie) :

merge(tt,z,by="row.names",all.x=TRUE)[,-(5:8)]

Now if you would work with dataframes instead of matrices, this would even become a whole lot easier :

z <- as.data.frame(z)
tt <- as.data.frame(tt)
merge(tt,z["symbol"],by="row.names",all.x=TRUE)

edited Sep 08 '14 at 11:32

answered May 17 '11 at 11:36

Joris Meys

106,551
31
221
263

Andrie · Accepted Answer · 2011-05-17T12:32:18.097

Use match to return your desired vector, then cbind it to your matrix

cbind(t, z[, "symbol"][match(rownames(t), rownames(z))])

             [,1]         [,2]         [,3]         [,4]   
GO.ID        "GO:0002009" "GO:0030334" "GO:0015674" NA     
LEVEL        "8"          "6"          "7"          NA     
Annotated    "342"        "343"        "350"        NA     
Significant  "1"          "1"          "1"          NA     
Expected     "0.07"       "0.07"       "0.07"       NA     
resultFisher "0.679"      "0.065"      "0.065"      NA     
ILMN_1652464 "0"          "0"          "1"          "PLAC8"
ILMN_1651838 "0"          "0"          "0"          "RND1" 
ILMN_1711311 "1"          "1"          "0"          NA     
ILMN_1653026 "0"          "0"          "0"          "GRA"

PS. Be warned that t is base R function that is used to transpose matrices. By creating a variable called t, it can lead to confusion in your downstream code.

Your answer is very useful thank you. The only thing is that my code does not give the rigth output. If i only take this: z[, "symbol"][match(rownames(t), rownames(z))] a factor is created with NA and symbols but when i do cbind, the symbol number are replaced by a rondom value. Do anyone know that is wrong? Thanks — Lisann, May 17 '11 at 11:19

score 3 · Answer 3 · answered May 17 '11 at 11:06

3

Not perfect but close:

newcol<-sapply(rownames(t), function(rn){z[match(rn, rownames(z)), 5]})
cbind(data.frame(t), newcol)

answered May 17 '11 at 11:06

Nick Sabbe

11,684
1
43
57

score 2 · Answer 4 · answered Dec 04 '14 at 13:43

cbind.fill <- function(x, y){
  xrn <- rownames(x)
  yrn <- rownames(y)
  rn <- union(xrn, yrn)
  xcn <- colnames(x)
  ycn <- colnames(y)
  if(is.null(xrn) | is.null(yrn) | is.null(xcn) | is.null(ycn)) 
    stop("NULL rownames or colnames")
  z <- matrix(NA, nrow=length(rn), ncol=length(xcn)+length(ycn))
  rownames(z) <- rn
  colnames(z) <- c(xcn, ycn)
  idx <- match(rn, xrn)
  z[!is.na(idx), 1:length(xcn)] <- x[na.omit(idx),]
  idy <- match(rn, yrn)
  z[!is.na(idy), length(xcn)+(1:length(ycn))] <- y[na.omit(idy),]
  return(z)
}

score 2 · Answer 5 · edited Oct 10 '17 at 17:25

2

you can wrap -Andrie answer into a generic function

mbind<-function(...){
 Reduce( function(x,y){cbind(x,y[match(row.names(x),row.names(y)),])}, list(...) )
}

Here, you can bind multiple frames with rownames as key

edited Oct 10 '17 at 17:25

R.S.

2,093
14
29

answered Jan 03 '17 at 10:23

Adam Lee Perelman

820
8
9

Merge or combine by rownames

5 Answers5

Linked

Related