[R] Why does R do this? (Duncan Murdoch)

Hamed Ha h@medh@@eli @ending from gm@il@com
Wed Jan 9 12:16:54 CET 2019


Well, your question looks quite interesting to me.

*which *function normally returns a non-negative *integer *and if fails
to find the case, then returns integer(0) that is an integer with the zero
length. Logically it returns the right answer. Then, doing an operation on
nothing is pointless, however in mathematics, the complementary of nothing
is everything. Then you do not expect R to return the entire integer set! ;)


If you still like to use which, I advise you write a little function like
which0() that inherits the properties of *which *and checks for the special
cases.


Hamed.






Message: 1
Date: Tue, 8 Jan 2019 06:54:43 -0500
From: Duncan Murdoch <murdoch.duncan using gmail.com>
To: Nick Wray <nicholas.wray using ntlworld.com>, r-help
        <r-help using r-project.org>
Subject: Re: [R] Why does R do this?
Message-ID: <8ed2949e-314f-ef92-1982-1de5e31bed99 using gmail.com>
Content-Type: text/plain; charset="utf-8"; Format="flowed"

On 08/01/2019 4:28 a.m., Nick Wray via R-help wrote:
> y<-c(1,2,3)
> z<-which(y>3)

At this point z is a vector with no entries in it.

> z
> y<-y[-z]

-z is the same vector.  So y[z] and y[-z] are the same.

> y
>
> In the work I'm doing I often have this situation and have to make sure
that I condition on z being non-zero as y is now numeric(0) rather than the
set c(1,2,3).  Why does R do this?  Wouldn't it be more sensible for R to
simply leave the host set unchanged if there are no elements to take out?

No, it wouldn't.  You asked for no entries, so you get no entries.

Follow Thierry's advice, and don't use which() unless you really need a
vector of indices, and are prepared for an empty one.

Duncan Murdoch

	[[alternative HTML version deleted]]



More information about the R-help mailing list