Collapse

[Euslaner]

Sorry for this post, but I am trying to collapse one data set to merge with another. From the data set I am using to collapse, I typed with the result:

. xcollapse pid3 - rewardsbased evangelical christiannationalism hispanic christiannationalismfactor - liberal,by(ppstaten) saving(d:\prrinationalismstate.dta)
error:
christiannationalism = (mean) christiannationalism
christiannationalism = (mean) christiannationalism
name conflict
r(198);

but in the dataset christiannationalism is a variable that I should be able to collapse:

fsum christiannationalism

Variable | N Mean SD Min Max
----------------------+---------------------------------------------
christiannationalism | 4008 3.06 1.08 1.00 4.00

And it only appears once.

What am I doing wrong? Any help would be appreciated,.

Ric Uslaner

[Jeff Pitblado (StataCorp)]

The variable christiannationalism is probably showing up in your varlist specification twice. I suspect it shows up somewhere in the variable range pid3 - rewardsbased or in christiannationalismfactor - liberal. Try removing the explicit reference to christiannationalism in you call to collapse.

[Euslaner]

It is inky there once:

. des christiannationalism,full

storage display value
variable name type format label variable label
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
christiannationalism
byte %8.0g Q42_C * Do you have a favorable or an unfavorable view of each of the following people o

[Nick Cox]

My guess is identical to that of Jeff Pitblado (StataCorp) The reason for the error message name conflict is that you are asking twice (or more) for the same output.

We're limited here by the lack of a data example, and some unexplained details, contrary to longstanding requests in the Statalist FAQ, but the mentions of xcollapse and fsum are probably red herrings for different reasons.

xcollapse from SSC is probably reacting just as collapse would. I can't test that on RIc/Eric Uslaner's data; it's just a guess,

The fact that fsum from SSC shows results for just one variable when supplied with just one variable name should not seem surprising and is not germane to the problem.

Here is how to remove duplicates. I use a silly example in which variable names are repeated.

. sysuse auto, clear
(1978 automobile data)

. collapse turn trunk t*, by(foreign)
error:
       turn = (mean) turn
       turn = (mean) turn
name conflict
r(198);

. unab myvars : turn trunk t*

. di "`myvars'"
turn trunk trunk turn

. local myvars : list uniq myvars

. di "`myvars'"
turn trunk

So what should work for Eric/Ric is something like

unab myvars : pid3 - rewardsbased evangelical christiannationalism hispanic christiannationalismfactor - liberal
local myvars : list uniq myvars
collapse `myvars' ,by(ppstaten) saving(d:\prrinationalismstate.dta)

Make sure that you run that all at once, not line by line from a do-file editor window.