Importing data using Foreign

It is because you don't know whether you want it or not.

It is a bit more obvious with integer indexing, as in color[race]: if race
is NA you don't know what color to put in, but the result should be the
same length as race.

With logical indices, the behaviour is a bit annoying, but ultimately
follows from the coercion rules: You might think that you could treat NA as
FALSE (& the subset() function does just that), but then you'd get the
problem that x[NA] would differ from x[as.integer(NA)] because NA is of
mode "logical", lowest in the coercion hierarchy.

-pd

On 26 Aug 2020, at 17:06 , Elham Daadmehr <e.daadmehr at gmail.com> wrote:

Thanks guys. but I'm a bit confused. the input is the first column

(z[,1] and z1[,1]).

How is it possible that a subset of a non-NA vector, contains NA?

On Wed, Aug 26, 2020 at 4:58 PM Eric Berger <ericjberger at gmail.com>

wrote:

Good point! :-)


On Wed, Aug 26, 2020 at 5:55 PM peter dalgaard <pdalgd at gmail.com> wrote:
Offhand, I suspect that the NAs are in the 8th column.

On 26 Aug 2020, at 10:57 , Elham Daadmehr <e.daadmehr at gmail.com>

wrote:

Hi all,

I have a simple problem. I get stuck in using the imported spss data

(.sav)

using "read.spss".
I imported data (z) without any problem. After importing, the first

column

doesn't contain any "NA". but when I choose a subset of it (like:
z[z[,8]=="11"|z[,8]=="12"|z[,8]=="14",]), lots of NA appears (even in

the

first column).

The (.sav) file is the output of Compustat (WRDS).

It is terrible, I can't find the mistake.

Thank you in advance for your help,
Elham

      [[alternative HTML version deleted]]