How to get the count of all observations in a table in Stata 13?

Question

I would like to count observations of a variable made of observations of other variables which fulfill logical criteria and use it in an assert statement in order to do some testing of data constraints, e.g.

generate myvar = (var1 == 0 & var2 > 0) | (var1 == 0 & var3b > 0)
assert magiccommand(myvar) == 0

where magiccommand should be the command or function I'm looking for (doesn't have to be a function).

I tried

count: help count doesn't provide any information on how to count observations. Linkage of information in general is very poor.
sum(): doesn't do the job, the graphical output displays the number of observations, but Stata's poor language concept doesn't allow programmatic retrieval of the value, nor does it provide any points of entry for information retrieval in the manual via help.
by: is insufficiently explained in general and seems to be designed too powerful.
I checked confirm conditional statement applies to >0 observations in Stata which is almost the same question as mine, but I'm not sure how I'm supposed to apply the answer to assert and still too shocked about the complexity of the command for such a ridiculously easy task.
help _n: "contains the number of the current observation" without a link to an explanation what the current observation is...

I'd be very interested in a way how I could have found out an answer myself because it's very hard to use closed source software which is built on selling expensive books compensating the malious manual.

I have to use Stata 13 on Windows 7.

What does "malious" mean? I think it's best not to mix specific questions with diatribes against software you don't like. Either you incite flames or people not interested in low-level debate are distracted from the question you want answered. Stata is indeed proprietary software you have to buy, just like Windows 7, but the entire .pdf documentation is now open source. — Nick Cox
@NickCox that is as interesting as it is problematic from the POV of contribution to the docs. I assume you mean thefullwiki.org/Stata/Documentation?! — Karl Richter
Goodness no! That site covers no more than a few bits and pieces and is of no serious value. Start with stata.com/features/documentation or within Stata go for .pdf documentation in the menu. — Nick Cox
@NickCox I saw that, but nothing except © Copyright 1996–2014 StataCorp LP, so what do you mean with the documentation being now open source? — Karl Richter
I mean anyone can read it. If you want to insist that anybody must be able to edit it or write it too, then it's not open source. See how well the open source idea has worked with the Wiki site you found. Even with R I don't think there is much attention to rewriting what anybody else wrote: people just document what they wrote and that's the same in Stata too, with a few thousand user-written packages. (By the way, you did try clicking on the links? The page is not just an advert; it's a portal to the manuals.) — Nick Cox

Roberto Ferrer Roberto Ferrer · Accepted Answer · 2014-11-16T22:55:13

You can use summarize:

clear
set more off

input ///
var1 var2 var3b
0 1 5
0 0 0
0 0 3
1 0 0
end

generate myvar = (var1 == 0 & var2 > 0) | (var1 == 0 & var3b > 0)

list

*-----

summarize myvar
assert r(sum) == 0

Another way is

<snip>    

count if myvar
assert r(N) == 0

Yet another:

<snip>

gen c = sum(myvar)
assert c[_N] == 0

help return is one reference. help subscripting is another one.

search count observations takes you to count, which is very clear on what it does, including examples. I really don't understand why you say it "doesn't provide any information on how to count observations". See also

The Stata Journal, Volume 7 Number 1: pp. 117-130, Speaking Stata: Making it count, by Nicholas J. Cox (http://www.stata-journal.com/article.html?article=pr0029).

With all due respect, I think you've done a really poor job of reading the help files and the manuals. You really don't explain what you mean by length() in your post. You assume that Stata's programming paradigm is equal to those of other languages, and because you haven't studied it, you assume it is a bad one. It's different, I agree. It does some things fine, others not so fine, but so do all other languages. Stata developers have made it possible to seek help in many ways (really). This is clearly explained in the very first sections of the User's Manual.

If I've given code that doesn't solve your problem, then you probably need to explain yourself better.

How to get the count of all observations in a table in Stata 13?

1 Answers