Skip to content

Impact of cex changing as a function of mfrow

5 messages · Dennis Fisher, Duncan Murdoch

Dennis Fisher
# 
R 2.15.1
OS X (MLion)

Colleagues,

I am aware that changes in mfrow / mfcol in par() affect cex (from help: In a layout with exactly two rows and columns the base value of ?"cex"? is reduced by a factor of 0.83: if there are three or more of either rows or columns, the reduction factor is 0.66).

I generate a multipage PDF in which mfrow varies such that cex is impacted.  This affect mtext in the outer margin.  Sample code is pasted at the bottom of this email.  The impact is most obvious if one examines the text at the bottom of each page as one moves page-to-page.

Does anyone have a suggestion for how to overcome this (other than using brute force).

Dennis

Dennis Fisher MD
P < (The "P Less Than" Company)
Phone: 1-866-PLessThan (1-866-753-7784)
Fax: 1-866-PLessThan (1-866-753-7784)
www.PLessThan.com

# In a layout with exactly two rows and columns the base value of ?"cex"? is reduced by a factor of 0.83: if 
# there are three or more of either rows or columns, the reduction factor is 0.66.

FINDCEX	<- function()
	{
	CORRECT	<- 1
	MFROW		<- par()$mfrow
	MFCOL		<- par()$mfcol
	TEST		<- all(MFROW == MFCOL)
	if (TEST && MFROW == c(2,2))				CORRECT	<- 1 / 0.83
	if (TEST && (MFROW[1] >= 3 | MFROW[2] >= 3))	CORRECT	<- 1 / 0.66
	if (!TEST)	cat("MFROW does not equal MFCOL\n")
	cat(MFROW, CORRECT, "\n")
	return(CORRECT)
	}
pdf("TestCEX.pdf", 8, 6)
par(mfrow=c(1,1), omi=c(1,1,1,1))
plot(1)
mtext(outer=T, side=1, line=0, cex=1, "line 0")
mtext(outer=T, side=1, line=FINDCEX(), cex=1, "line FINDCEX")
mtext(outer=T, side=3, line=0, cex=1, "line 0")
mtext(outer=T, side=3, line=1, cex=1, "line 1")
mtext(outer=T, side=2, line=0, cex=1, "line 0")
mtext(outer=T, side=2, line=FINDCEX(), cex=1, "line FINDCEX")
mtext(outer=T, side=4, line=0, cex=1, "line 0")
mtext(outer=T, side=4, line=1, cex=1, "line 1")


par(mfrow=c(1,2), omi=c(1,1,1,1))
plot(1)
mtext(outer=T, side=1, line=0, cex=1, "line 0")
mtext(outer=T, side=1, line=FINDCEX(), cex=1, "line FINDCEX")
mtext(outer=T, side=3, line=0, cex=1, "line 0")
mtext(outer=T, side=3, line=1, cex=1, "line 1")
mtext(outer=T, side=2, line=0, cex=1, "line 0")
mtext(outer=T, side=2, line=FINDCEX(), cex=1, "line FINDCEX")
mtext(outer=T, side=4, line=0, cex=1, "line 0")
mtext(outer=T, side=4, line=1, cex=1, "line 1")

par(mfrow=c(2,2), omi=c(1,1,1,1))
plot(1)
mtext(outer=T, side=1, line=0, cex=1, "line 0")
mtext(outer=T, side=1, line=FINDCEX(), cex=1, "line FINDCEX")
mtext(outer=T, side=3, line=0, cex=1, "line 0")
mtext(outer=T, side=3, line=1, cex=1, "line 1")
mtext(outer=T, side=2, line=0, cex=1, "line 0")
mtext(outer=T, side=2, line=FINDCEX(), cex=1, "line FINDCEX")
mtext(outer=T, side=4, line=0, cex=1, "line 0")
mtext(outer=T, side=4, line=1, cex=1, "line 1")

par(mfrow=c(3,3), omi=c(1,1,1,1))
plot(1)
mtext(outer=T, side=1, line=0, cex=1, "line 0")
mtext(outer=T, side=1, line=FINDCEX(), cex=1, "line FINDCEX")
mtext(outer=T, side=3, line=0, cex=1, "line 0")
mtext(outer=T, side=3, line=1, cex=1, "line 1")
mtext(outer=T, side=2, line=0, cex=1, "line 0")
mtext(outer=T, side=2, line=FINDCEX(), cex=1, "line FINDCEX")
mtext(outer=T, side=4, line=0, cex=1, "line 0")
mtext(outer=T, side=4, line=1, cex=1, "line 1")

graphics.off()
system("open TestCEX.pdf")
Duncan Murdoch
#
On 12-09-02 1:40 PM, Dennis Fisher wrote:
I am not sure what's wrong with brute force here.  You know the formula 
for the reduction; just apply a corresponding increase first.

Duncan Murdoch
Dennis Fisher
# 
Duncan

Perhaps I did not explain sufficiently -- the code that I sent makes the correction in FINDCEX.  But, that correction does not accomplish the intended goal -- as evidenced by the graphic that is created from the code that I sent.

The option of "brute force" would require trial and error.  I was hoping that invoking some other option in mtext, such as padj, would accomplish my goal of needing only to multiply the line= value by the value returned from my function.

Dennis

Dennis Fisher MD
P < (The "P Less Than" Company)
Phone: 1-866-PLessThan (1-866-753-7784)
Fax: 1-866-PLessThan (1-866-753-7784)
www.PLessThan.com
On Sep 2, 2012, at 11:43 AM, Duncan Murdoch <murdoch.duncan at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Duncan Murdoch
    

    

      
      #
    

  


  

      
On 12-09-02 3:52 PM, Dennis Fisher wrote:

            

      
      
      
    

        
You're right, I didn't understand you.

Here's what I see in your sample code display (in 2.15.0 on a Mac with 
OSX 10.6.8):

The spacing between "line 0" and "line 1" (on the top and right) gets 
smaller from page 1 to page 4.

The size of the margin text is the same in all pages.

The size of "Index" within the subfigures is smaller in the later pages.

There are some other small movements of text from page to page, but not 
much.

What do you see?

Duncan Murdoch

            

      
      
      
    

        

      
      
    

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Dennis Fisher
    

    

      
      #
    

  


  

      
Duncan

I are getting closer to an understanding.  
	1.  "index" is not relevant to the discussion.  i am focused solely on margin text
	2.  the size pf the margin text does not change between pages
	3.  however, the location does change, moreso on the top / right than on the left / bottom.

I rewrote the code to better illustrate the problem:
###########################
FINDCEX <- function()
        {
        CORRECT <- 1
        MFROW           <- par()$mfrow
        MFCOL           <- par()$mfcol
        TEST            <- all(MFROW == MFCOL)
        if (TEST && MFROW == c(2,2))                            CORRECT <- 1 / 0.83
        if (TEST && (MFROW[1] >= 3 | MFROW[2] >= 3))    CORRECT <- 1 / 0.66
        if (!TEST)      cat("MFROW does not equal MFCOL\n")
        cat(MFROW, CORRECT, "\n")
        return(CORRECT)
        }
DRAW    <- function(N)
        {
        for (i in 1:N)
                {
                plot(1, bty="n")
                box(which="inner")
                }
        for (SIDE in 1:2)
                {
                mtext(outer=T, side=SIDE, line=0, cex=1, "line 0")
                mtext(outer=T, side=SIDE, line=FINDCEX(), cex=1, "line FINDCEX")
                mtext(outer=T, side=SIDE, line=-FINDCEX(), cex=1, "line FINDCEX")
                }
        for (SIDE in 3:4)
                {
                mtext(outer=T, side=SIDE, line=0, cex=1, "line 0")
                mtext(outer=T, side=SIDE, line=1, cex=1, "line 1")
                }
        }
pdf("TestCEX.pdf", 10, 8)
par(mfrow=c(1,1), omi=c(1,1,1,1))
DRAW(1)
par(mfrow=c(2,2), omi=c(1,1,1,1))
DRAW(4)
par(mfrow=c(3,3), omi=c(1,1,1,1))
DRAW(9)
graphics.off()
system("open TestCEX.pdf")

system("open TestCEX.pdf")	## in Windows, "start" instead of "open"
################

If you execute this code, you will see:
1.  the box at the outer border is identical on all pages (as expected)
2.  the margin side on the left side superimposes between pages
3.  on the bottom, margin text is spaced the same on each page (unlike at the top).  However, the vertical position differs between pages (most obvious on comparing pages 1 and 3).  

I could correct the vertical positioning by brute force, i.e., adding a constant to each mtext command (applying a different value based on the value of mfrow) -- It turn out that a correction of 0.2 (for mfrow=c(2,2) and 0.4 (for mfrow=c(2,2) aligns everything properly.  .  However, this is not elegant.  The correction factor took care of line-to-line spacing but not vertical position.

Dennis



Dennis Fisher MD
P < (The "P Less Than" Company)
Phone: 1-866-PLessThan (1-866-753-7784)
Fax: 1-866-PLessThan (1-866-753-7784)
www.PLessThan.com

        
On Sep 2, 2012, at 10:40 AM, Dennis Fisher wrote: