Looks to me as if you've been bytten by something.
Etiquette says that I should not claim it to be
a bug.
. whelp datatypes
makes it clear that a byte
can hold non-missing values between -127 and
100.
Can you
. l if !inlist(hadPCI,0,1)
Nick
[email protected]
Jeph Herrin
> I've been using Stata a long time and never seen this.
>
> . gen byte PCI=(initrs==4|initrs==5|initrs==6)
> . gen byte hadPCI=PCI
> . tab hadPCI
>
> hadPCI | Freq. Percent Cum.
> ------------+-----------------------------------
> -128 | 2 0.00 0.00
> 0 | 932,238 91.80 91.80
> 1 | 83,218 8.20 100.00
> ------------+-----------------------------------
> Total | 1,015,458 100.00
>
> Where does that -128 come from?
>
> I am using Stata 8, latest updates, and I have run into this
> problem with *three* different byte variables I generate using
> this dataset; the only distinctive thing about the dataset is
> it's size (>500MB). The problem is difficult to reproduce, as
> well, so it's hard to figure out what is going on; for instance,
> if I add a line that tabs the first variable (PCI) then there
> is no trouble. I did try it without the "byte" designations, and
> got the same spurious value.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
