Thanks to Austin's and then Vince's clarification, here is an
example of how to get -bargap()- to do something in -catplot-.
. sysuse auto, clear
. catplot bar foreign rep78, asyvars bargap(-30)
bw: I pointed out earlier that -catplot, over()- is legal, because the
extra options you can specify don't exclude -over()-, but it is _not_
the intended use of -catplot- and in some circumstances it won't give
the correct numerical results as far as the user is concerned. Vince
repeated that comment. Here is the warning for a third time!
Something like
. catplot bar ind_methd year if ind_methd !="emerg", asyvars
or
. catplot bar year ind_methd if ind_methd !="emerg", asyvars
is better practice.
(-catplot- is on SSC.)
Nick
[email protected]
b. water
austin, thanks for your input and clarification.
i tried (FWIW) to get travel.dta but there must an internal check
mechanism either at statacorp or my end that blocked its use on my stata
8.2. btw you are right that the [G] i referred to was of stata 8. am not
quite sure whether i can get travel.dta with stata 9 (a colleague has
it) so that i can see what you are trying to illustrate. conceptually
though i think i understand what you are trying to say wrt your example
as well as the one on my stata 8 [G] p. 125 and gave me an idea to try.
i tried this:
obs year ind_methd rapid standard
4 2002 emerg
5 2002 emerg
14 2002 rapid rapid
1 2003 rapid rapid
8 2003 rapid rapid
3 2004 emerg
2 2005 standard standard
6 2005 rapid rapid
10 2005 rapid rapid
11 2005 standard standard
12 2005 standard standard
7 2006 rapid rapid
9 2006 standard standard
13 2006 rapid rapid
then:
. graph bar rapid standard, over(year)
type mismatch
r(109);
. catplot bar rapid standard, over(year)
no observations
r(2000);
BUT:
if i fill the blanks with "." i.e. inserting . (dot or period) to the
blank observations in both rapid and standard, i nearly get to where i
wanted:
obs year ind_methd rapid standard
4 2002 emerg . .
5 2002 emerg . .
14 2002 rapid rapid .
1 2003 rapid rapid .
8 2003 rapid rapid .
3 2004 emerg . .
2 2005 standard . standard
6 2005 rapid rapid .
10 2005 rapid rapid .
11 2005 standard . standard
12 2005 standard . standard
7 2006 rapid rapid .
9 2006 standard . standard
13 2006 rapid rapid .
. graph bar rapid standard, over(year) -> still returns
type mismatch
r(109);
however,
. catplot bar rapid standard, over(year) bargap (-30)
come tantalizingly close to what i wanted if only i can suppress the "."
being displayed or used in the resulting graph.
would therefore appreciate further advice/help if possible. otherwise i
will settle for:
. catplot bar ind_methd if ind_methd !="emerg", over(year) -> in this
case -bargap- option would not be applicable as per austin's
explanation.
regards,
bw
> Date: Thu, 10 Jan 2008 10:27:20 -0500
> From: [email protected]
> To: [email protected]
> Subject: Re: st: producing graph bar help
>
> Dear Nick and "b. water":
> In his first post "b. water" announces Stata 8.2 and cites [G] p.125
> (presumably the Stata 8 manual). The negative bargap option is
> illustrated with
>
> webuse citytemp, clear
> gr bar tempjuly tempjan, over(div) over(reg) nofill bargap(-30)
>
> in the Stata 10 manual [G] p.62, and it may not be clear, but the
> bargap controls the relative plot position of tempjuly and tempjan,
> not the positions of those two variables' bars with respect to over()
> categories. Compare the following:
>
> webuse travel, clear
> drop if choice==0
> g x=1
> replace party=3 if party>2
> gr bar (count) x, over(mode) over(party) nofill bargap(-30)
> separate x, by(mode)
> gr bar (count) x1 (count) x2 (count) x3 (count) x4, over(party)
bargap(-30)
> collapse (count) x, by(mode party)
> forv i=1/4 {
> g `: label (mode) `i''=x if mode==`i'
> }
> gr bar air train bus car, over(party) bargap(-30)
*
* 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/
