Skip to content

Convert matrix to numeric

6 messages · Jeffrey Joh, Duncan Murdoch, Kenneth Takagi +3 more

#
I have a matrix that looks like this:


structure(c("0.0376673981759913", "0.111066500741386", "1", "1103", 
"18", "OPEN", "DEPR", "0.0404073656092023", "0.115186044704599", 
"1", "719", "18", "OPEN", "DEPR", "0.0665342096693433", "0.197570061769498", 
"1", "1103", "18", "OPEN", "DEPR", "0.119287147905722", "0.356427096010845", 
"1", "1103", "18", "OPEN", "DEPR"), .Dim = c(7L, 4L), .Dimnames = list(
    c("Sn", "SlnC", "housenum", "date", "hour", "flue", "pressurization"
    ), c("10019.BLO", "1002.BLO", "10020.BLO", "10021.BLO")))

 

How do I convert rows 1-5 to numeric?  I tried mode() <- "numeric" but that doesn't change anything.

 

I also tried converting this to a table then converting to numeric, but I got: (list) object cannot be coerced to type 'double'

 

Jeff
#
On 03/08/2011 3:04 PM, Jeffrey Joh wrote:
Every entry in a matrix has the same type, so you can't change just 
those rows other than by extracting them into a separate matrix and 
changing that.

Duncan Murdoch
#
How about
Matrix[1:5,]=as.numeric(Matrix[1:5,])
-Ken Hutchison
On Aug 3, 2554 BE, at 3:04 PM, Jeffrey Joh <johjeffrey at hotmail.com> wrote:

            
#
Hi Jeffrey,
On Wed, Aug 3, 2011 at 3:04 PM, Jeffrey Joh <johjeffrey at hotmail.com> wrote:
Thank you for providing a small working example.
Two things are going on here. First, a matrix can only contain one kind of data.
For this example, since there are strings the whole thing has to be character.
A data frame is intended to hold different kinds of data, but each column has
to be a single type. So if you want those values to be numeric instead of
character, you'll need to transpose your matrix and convert it to a data frame.

tempdata <- structure(c("0.0376673981759913", "0.111066500741386", "1", "1103",
"18", "OPEN", "DEPR", "0.0404073656092023", "0.115186044704599",
"1", "719", "18", "OPEN", "DEPR", "0.0665342096693433", "0.197570061769498",
"1", "1103", "18", "OPEN", "DEPR", "0.119287147905722", "0.356427096010845",
"1", "1103", "18", "OPEN", "DEPR"), .Dim = c(7L, 4L), .Dimnames = list(
   c("Sn", "SlnC", "housenum", "date", "hour", "flue", "pressurization"
   ), c("10019.BLO", "1002.BLO", "10020.BLO", "10021.BLO")))

tempdata <- data.frame(t(tempdata), stringsAsFactors=FALSE)

Once you have the right kind of object, you can convert the five columns
of interest to numeric. This needs to be done a column at a time, I think:
tempdata[, 1:5] <- apply(tempdata[,1:5], 2, as.numeric)

Sarah
#
Ken,

You can't store the numeric values back in the matrix, because rows 6 and 7 contain character values.  Everything will just be converted back to character.  You need to create a new matrix for the numeric values.

Hope this is helpful,

Dan

Daniel J. Nordlund
Washington State Department of Social and Health Services
Planning, Performance, and Accountability
Research and Data Analysis Division
Olympia, WA 98504-5204
#
Here's what you _should_  do

1) transpose
2a) as.data.frame
3a) fix the stupid default stringsAsFactor behavior
4a) convert the first 5 columns to numeric

dfrm <- as.data.frame( t( structure(.) ) )
dfrm[, 1:5] <-lapply(dfrm[, 1:5], as.character)
dfrm[, 1:5] <-lapply(dfrm[, 1:5], as.numeric)

Or:
1) transpose
2b) as.data.frame with stringsAsFactors= FALSE
3b) convert to numeric
On Aug 3, 2011, at 3:04 PM, Jeffrey Joh wrote:

            
David Winsemius, MD
West Hartford, CT