[R] switch of cex adjustment with mfrow?

[Jannis]

Thanks guys for your help. Your suggestions will work in my case, the behaviour beeing dependent on the order of arguments (which Steve mentions) is however, not the way I would expect R to work. In addition, the way I interpret the documentation:
"A numerical value giving the amount by which plotting text and symbols should be magnified relative to the default. This starts as 1 when a device is opened, and is reset when the layout is changed, e.g. by setting mfrow."
is also different from the behaviour you describe. Before posting, I read this in a way that setting mfrow always changes the cex value (either the default value or the one you supply) later, so that setting:
par(mfrow=c(2,2), cex = 1)
would still produce a cex <1 later so that you have to run:
par(mfrow=c(2,2), cex = 1/0.66)
to get a effetive cex of one (after the mentioned reset). Maybe it would be good to make the documentation a bit more specific here?

Best 
Jannis 

    S Ellison <S.Ellison at LGCGroup.com> schrieb am 12:31 Donnerstag, 3.August 2017:
 

 > use
> 
> par(mfrow=c(2,2), cex = 1)

This does work as written. But when I first checked single-call setting, an  mfrow change to cex in the same call superseded cex=1; hence my suggestion to use separate calls to par().
Further checking confirms that the result of a call to par is dependent on argument specification order in the call:

par(mfrow=c(2,2), cex = 1)
par("cex")
    # 1

par(cex=1, mfrow=c(2,2))
par("cex")
    # 0.83

This obviously isn't a problem, though as an aside I didn't immediately find any comment on the effect of argument order in ?par. It just means you have to be careful exactly what you specify.
It may also mean that for future-proofing against possible changes in par execution order, you might want to use separate calls anyway.

Steve Ellison


*******************************************************************
This email and any attachments are confidential. Any use...{{dropped:12}}