[R] reading in a tricky computer program output

Gabor Grothendieck ggrothendieck at gmail.com
Sun Feb 5 16:29:18 CET 2006


Its not clear to me what format you want to put the data in but this
will read it into a list, one list element per lower triangular matrix.
Modify to suit.

DF <- read.table("myfile.dat", fill = TRUE)
id <- cumsum(is.na(DF[,2]))
result <- by(DF, id, as.matrix)

# if the input is in the second format add this line after the above
result2 <- lapply(result, function(x) rbind(NA, x))


On 2/5/06, Taka Matzmoto <sell_mirage_ne at hotmail.com> wrote:
> Hi R user
>
> I need to read in some values from a computer program output.
>
> I can't change the output format because the developer of the program
> doesn't allow to change the format of output.
>
> There are two formats.
>
> First one looks like this
>
> if I have 10 variables,
>
> ------------------------------------------------------------------------------------------------------
>       [ 1]              [2]           [3]              [4]            [5]
> [ 1]  0.000
> [ 2]  0.001         0.000
> [ 3] -0.002         0.019         0.000
> [ 4]  0.012        -0.004        -0.020         0.000
> [ 5] -0.015         0.003         0.011         0.008         0.000
> [ 6]  0.005        -0.008        -0.005         0.002         0.005
> [ 7]  0.008        -0.007         0.013         0.003         0.007
> [ 8] -0.014        -0.011        -0.010        -0.025         0.002
> [ 9]  0.006         0.003        -0.010         0.002        -0.020
> [10] 0.006         0.010        -0.006         0.005         0.008
> [ 6]  0.000
> [ 7] -0.037         0.000
> [ 8]  0.010         0.027         0.000
> [ 9]  0.032        -0.004         0.008         0.000
> [10] -0.008        -0.011         0.015        -0.020         0.000
>
> ------------------------------------------------------------------------------------------------
> NOTE: I put [number] to show that this output is similar to a lower diagonal
> matrix including diagonal. In an ouput there is no [number]
>
>
> The second format looks like this
> --------------------------------------------------------------------------------------
>       [1]              [2]             [3]           [4]              [5]
> [ 2] -0.002
> [ 3]  0.003        -0.053
> [ 4] -0.026         0.010         0.045
> [ 5]  0.023        -0.008        -0.025        -0.016
> [ 6] -0.012         0.023         0.013        -0.005        -0.011
> [ 7] -0.031         0.031        -0.054        -0.013        -0.027
> [ 8]  0.040         0.042         0.031         0.075        -0.007
> [ 9] -0.012        -0.009         0.023        -0.005         0.037
> [10] -0.013        -0.027         0.014        -0.013        -0.020
> [ 7]  0.127
> [ 8] -0.035        -0.166
> [ 9] -0.083         0.015        -0.027
> [10]  0.021         0.047        -0.052         0.048
> ---------------------------------------------------------------------------------------------------------
> NOTE: I put [number] to show that this output is similar to a lower diagonal
> matrix without diagonal. In an ouput there is no [number]
>
> The problem of this format is the fixed column length ( 5 columns)
>
> To make matter worse, the number of variables keep changing (10, 20, 30, 40,
> 50, 60,70,80,90, and 100) so I need to take into the number of variables
> when I write a R function to read in these numbers.
>
> If the number of variables is 80, the output is very long.
>
> I only came up with this tedious one.
>
> First I read in the output using scan() and then make it a numeric vector
>
> I created 10 character vectors. Creating a 100 variable character vector is
> the most boring things
>
> I have ever done.
>
> one of the character vectors that matchs with the first 10 variable output
> is
>
> first.10<-c(
>            "i.001.001",
>            "i.002.001","i.002.002",
>            "i.003.001","i.003.002","i.003.003",
>            "i.004.001","i.004.002","i.004.003","i.004.004",
>            "i.005.001","i.005.002","i.005.003","i.005.004","i.005.005",
>            "i.006.001","i.006.002","i.006.003","i.006.004","i.006.005",
>            "i.007.001","i.007.002","i.007.003","i.007.004","i.007.005",
>            "i.008.001","i.008.002","i.008.003","i.008.004","i.008.005",
>            "i.009.001","i.009.002","i.009.003","i.009.004","i.009.005",
>            "i.010.001","i.010.002","i.010.003","i.010.004","i.010.005",
>            "i.006.006",
>            "i.007.006","i.007.007",
>            "i.008.006","i.008.007","i.008.008",
>            "i.009.006","i.009.007","i.009.008","i.009.009",
>            "i.010.006","i.010.007","i.010.008","i.010.009","i.010.010"
>           )
>
> one of the character vectors that matchs with the second 10 variable output
> is
>
> second.10<-c(
>            "i.002.001",
>            "i.003.001","i.003.002",
>            "i.004.001","i.004.002","i.004.003",
>            "i.005.001","i.005.002","i.005.003","i.005.004",
>            "i.006.001","i.006.002","i.006.003","i.006.004","i.006.005",
>            "i.007.001","i.007.002","i.007.003","i.007.004","i.007.005",
>            "i.008.001","i.008.002","i.008.003","i.008.004","i.008.005",
>            "i.009.001","i.009.002","i.009.003","i.009.004","i.009.005",
>            "i.010.001","i.010.002","i.010.003","i.010.004","i.010.005",
>            "i.007.006",
>            "i.008.006","i.008.007",
>            "i.009.006","i.009.007","i.009.008",
>            "i.010.006","i.010.007","i.010.008","i.010.009"
>           )
>
> and then assign the character vector to the numeric vector by
>
> names<-first.10
> first.10 = numeric.vector
> combined.one <- cbind(names,first.10)
> container <- diag(10)
> for (i in 1:(10*10))
>    {
>        k   <- as.numeric(substr(combined.one[i,1],7,9))
>        l   <- as.numeric(substr(combined.one [i,1],3,5))
>        val <- as.numeric(combined.one [i,2])
>        container [k,l] <- val
>    }
>
> container <- t(container )
>
> Is there any other neat way to do this?
>
> Any help would be appreciated
>
> TM
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
>




More information about the R-help mailing list