R help - Mar 2015 - Using split.screen with a definition matrix of the screens seems to be a mess. I don't understand!

I have a particular need to divide the device space to draw different plots
and texts, so I decided to use split.screen using a matrix to define the
different space partitions. 

My code and explanation is as follows:

# ------ START OF R CODE ----

dirGraf <- "TEST/" # A directory to put the result

nc <- 4 # number of data screens

tit <- "My title"
righttext <- "right text"
lefttext <- "left text"

# To ensure there are no graphics devices on:
graphics.off()

# I want to create 7 plotting windows (screens) with the following 
# arrangement
# I'm supposing the NDC coordinates are such that the lower-left corner
# is at (0,0) and the upper-right corner is at (1,1). Am I right?
#
#       +-----------------------------------+ (1,1)
#       |                 5    h=0.5/5.5    |
#       +--+-----------------------------+--+
#       |  |              1    h=1.0/5.5 |  |
#       |  +-----------------------------+  |
#       |  |              2    h=1.0/5.5 |  |
#       |  +-----------------------------+  |
#       |  |              3    h=1.0/5.5 |  |
#       |6 +-----------------------------+ 7|
#       |  |                             |  |
#       |  |              4    h=2.0/5.5 |  |
#       |  |                             |  |
#       +--+-----------------------------+--+
#    (0,0)
#
# I want 4 information (central) windows, from 1 to 4. However these windows
# must vary in their heights, because window 4 is twice the 
# height the others.
# I want an upper window to put a title to the whole plot, and two 
# lateral windows to add some texts.
#
# So I will build a four-column matrix, as instructed by the split.screen 
# documentation. Each column correspond in order to the following 
# characteristics, of each window (represented by each row), namely:
# 1. left, 2. right, 3. bottom, 4. top.
#

# First I will create a vector with the first four windows heights
Yinc <- c(2, rep(1,3))/5.5 
# To have the NDC device coordinates (?) I will accumulate:
Yinc <- Reduce('+', Yinc, accumulate=T)
# To put this information in the right window order I invert the vector
Yinc <- Yinc[4:1] 
# Then I will build the matrix for the first four windows:
Mm <- cbind(left=1/6, right=5/6, bottom=c(Yinc[2:4],0), top=Yinc)
# Now, let's add upper and lateral windows (screens 5 to 7)
Mm <- rbind(Mm, 
            # +--------+--------+--------+--------+
            # | left   | right  | bottom |   top  |
            # +--------+--------+--------+--------+
            c(   0    ,   1    , Yinc[1],    1   ),
            c(   0    ,  1/6   ,   0    , Yinc[1]),
            c(  5/6   ,   1    ,   0    , Yinc[1])
)

# margins for most of the cases
gpar <- list(mar=c(0,2.1,0,0))

# the device name (a pdf file)
gname <- paste0(dirGraf, "Test.pdf") 

# Test table to plot, just a line from (0,0) to (1,1)
tt <- data.frame(x=c(0,1), y=c(0,1))
# Let's open the device:
pdf(gname, width=7, height=9.11) 
# Let's split the device space (I think this is the default, isn't it?)
split.screen(Mm)

# Let's put a title in the upper window
screen(5)
par(gpar)
plot(c(0,1), c(0,1), ylab="", axes=F, type="n")
text(0.5, 0.5, tit, cex=1.5)

# For each "data" window
for (ii in 1:nc) { 
    screen(ii) # select screen 1 to 4
    par(gpar)
 
    plot(tt, ylab=LETTERS[ii], xlab="", type="l",
xaxt="n")
}
# A text in the left window
screen(6)
par(mar=c(0.1,0.1,0.1,0.1))
plot(c(0,1), c(0,1), axes=F, type="n")
text(0.5, 0.5, lefttext, srt=90, cex=1.2)
# A text in the right window
screen(7)
par(mar=c(0.1,0.1,0.1,0.1))
plot(c(0,1), c(0,1), axes=F, type="n")
text(0.5, 0.5, righttext, srt=90, cex=1.2)
# close graphics
graphics.off()

# ------ END OF R CODE ----------------

At the end I have a totally messy pdf file with the four main windows
at the top and with no difference in their heights, and the other windows
in any place R (not me) chose to put them.

Furthermore the R interpreter issues a message:

 "In par(new = TRUE) : call par(new=TRUE) without graphic"

Do you have any comments on this?

Thanks a lot!
  -Sergio.

Hi Sergio,
In order to get this configuration you will have to do multiple splits. Try
this to see how you get what you want, and notice that screens 2 and 4 are
now unusable as they have been split to make more screens.

split.screen(figs=matrix(c(0,1,0.91,1,0,1,0,0.91),nrow=2,byrow=TRUE))
split.screen(figs=matrix(c(0,0.09,0,1,0.09,0.91,0,1,0.91,1,0,1),nrow=3,
 byrow=TRUE),screen=2)
split.screen(figs=matrix(c(0,1,0.8,1,0,1,0.6,0.8,0,1,0.4,0.6,0,1,0,0.4),
 nrow=4,byrow=TRUE),screen=4)
screen(1)
par(mar=c(0,0,0,0))
plot(1:2,type="n",axes=FALSE)
box()
text(1.5,1.5,1)
screen(3)
par(mar=c(0,0,0,0))
plot(1:2,type="n",axes=FALSE)
box()
text(1.5,1.5,3)
screen(5)
par(mar=c(0,0,0,0))
plot(1:2,type="n",axes=FALSE)
box()
text(1.5,1.5,5)
screen(6)
par(mar=c(0,0,0,0))
plot(1:2,type="n",axes=FALSE)
box()
text(1.5,1.5,6)
screen(7)
par(mar=c(0,0,0,0))
plot(1:2,type="n",axes=FALSE)
box()
text(1.5,1.5,7)
screen(8)
par(mar=c(0,0,0,0))
plot(1:2,type="n",axes=FALSE)
box()
text(1.5,1.5,8)
screen(9)
par(mar=c(0,0,0,0))
plot(1:2,type="n",axes=FALSE)
box()
text(1.5,1.5,9)

Jim


On Wed, Mar 18, 2015 at 12:41 PM, Julio Sergio Santana <
juliosergio at gmail.com> wrote:
> I have a particular need to divide the device space to draw different plots
> and texts, so I decided to use split.screen using a matrix to define the
> different space partitions.
>
> My code and explanation is as follows:
>
> # ------ START OF R CODE ----
>
> dirGraf <- "TEST/" # A directory to put the result
>
> nc <- 4 # number of data screens
>
> tit <- "My title"
> righttext <- "right text"
> lefttext <- "left text"
>
> # To ensure there are no graphics devices on:
> graphics.off()
>
> # I want to create 7 plotting windows (screens) with the following
> # arrangement
> # I'm supposing the NDC coordinates are such that the lower-left corner
> # is at (0,0) and the upper-right corner is at (1,1). Am I right?
> #
> #       +-----------------------------------+ (1,1)
> #       |                 5    h=0.5/5.5    |
> #       +--+-----------------------------+--+
> #       |  |              1    h=1.0/5.5 |  |
> #       |  +-----------------------------+  |
> #       |  |              2    h=1.0/5.5 |  |
> #       |  +-----------------------------+  |
> #       |  |              3    h=1.0/5.5 |  |
> #       |6 +-----------------------------+ 7|
> #       |  |                             |  |
> #       |  |              4    h=2.0/5.5 |  |
> #       |  |                             |  |
> #       +--+-----------------------------+--+
> #    (0,0)
> #
> # I want 4 information (central) windows, from 1 to 4. However these
> windows
> # must vary in their heights, because window 4 is twice the
> # height the others.
> # I want an upper window to put a title to the whole plot, and two
> # lateral windows to add some texts.
> #
> # So I will build a four-column matrix, as instructed by the split.screen
> # documentation. Each column correspond in order to the following
> # characteristics, of each window (represented by each row), namely:
> # 1. left, 2. right, 3. bottom, 4. top.
> #
>
> # First I will create a vector with the first four windows heights
> Yinc <- c(2, rep(1,3))/5.5
> # To have the NDC device coordinates (?) I will accumulate:
> Yinc <- Reduce('+', Yinc, accumulate=T)
> # To put this information in the right window order I invert the vector
> Yinc <- Yinc[4:1]
> # Then I will build the matrix for the first four windows:
> Mm <- cbind(left=1/6, right=5/6, bottom=c(Yinc[2:4],0), top=Yinc)
> # Now, let's add upper and lateral windows (screens 5 to 7)
> Mm <- rbind(Mm,
>             # +--------+--------+--------+--------+
>             # | left   | right  | bottom |   top  |
>             # +--------+--------+--------+--------+
>             c(   0    ,   1    , Yinc[1],    1   ),
>             c(   0    ,  1/6   ,   0    , Yinc[1]),
>             c(  5/6   ,   1    ,   0    , Yinc[1])
> )
>
> # margins for most of the cases
> gpar <- list(mar=c(0,2.1,0,0))
>
> # the device name (a pdf file)
> gname <- paste0(dirGraf, "Test.pdf")
>
> # Test table to plot, just a line from (0,0) to (1,1)
> tt <- data.frame(x=c(0,1), y=c(0,1))
> # Let's open the device:
> pdf(gname, width=7, height=9.11)
> # Let's split the device space (I think this is the default, isn't
it?)
> split.screen(Mm)
>
> # Let's put a title in the upper window
> screen(5)
> par(gpar)
> plot(c(0,1), c(0,1), ylab="", axes=F, type="n")
> text(0.5, 0.5, tit, cex=1.5)
>
> # For each "data" window
> for (ii in 1:nc) {
>     screen(ii) # select screen 1 to 4
>     par(gpar)
>
>     plot(tt, ylab=LETTERS[ii], xlab="", type="l",
xaxt="n")
> }
> # A text in the left window
> screen(6)
> par(mar=c(0.1,0.1,0.1,0.1))
> plot(c(0,1), c(0,1), axes=F, type="n")
> text(0.5, 0.5, lefttext, srt=90, cex=1.2)
> # A text in the right window
> screen(7)
> par(mar=c(0.1,0.1,0.1,0.1))
> plot(c(0,1), c(0,1), axes=F, type="n")
> text(0.5, 0.5, righttext, srt=90, cex=1.2)
> # close graphics
> graphics.off()
>
> # ------ END OF R CODE ----------------
>
> At the end I have a totally messy pdf file with the four main windows
> at the top and with no difference in their heights, and the other windows
> in any place R (not me) chose to put them.
>
> Furthermore the R interpreter issues a message:
>
>  "In par(new = TRUE) : call par(new=TRUE) without graphic"
>
> Do you have any comments on this?
>
> Thanks a lot!
>   -Sergio.
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]