<>
I would say that the "mistake" that you made is an obvious hazard, as the
-sum- -total- and -count- functions for -egen- are not easily separated in
users` heads, and the added complication of an -expression- being allowed as
the argument compounds the problems.
Admittedly, I had to read the definition in the -h egen, count()- to solve
your puzzle. Still, you can combine the -count()- function with the -by-
prefix, and it can be handy then to see how many non-missing entries there
are for a certain variable...
*************
clear*
set obs 5
gen id=_n
expand 10
set seed 43987
gen x=rnormal()
//25 % artificially missing
replace x=. if /*
*/ runiform()<0.25
//How many non-missing per group?
bys id: egen mycount=/*
*/ count(x)
list, noobs sepby(id)
*************
HTH
Martin
-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von [email protected]
Gesendet: Donnerstag, 13. August 2009 11:16
An: statalist
Betreff: Re: st: AW: AW: Something wrong with the -count- option in the
-egen-?
Martin，
Thank you for your help, and now I know the inner rationale for -count- in
the -egen-.
The -total- option is helpful.
BTW, what is the most value for -count- option in the -egen- ?
I can get r(N) quickly after stand alone -count- .
Best regards,
Rose.
----- Original Message -----
From: Martin Weiss <[email protected]>
To: <[email protected]>
Subject: st: AW: AW: Something wrong with the -count- option in the -egen-?
Date: 2009-8-13 17:02:02
<>
BTW, the standalone -count- command leaves "r(N)" behind, so you can process
the number further. If you want it in your dataset, you can use -egen,
total()-
*************
clear
input x
1
2
3
5
.
end
egen yyy=/*
*/ total(x==2)
list, noobs
*************
HTH
Martin
-----Urspr?gliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Martin Weiss
Gesendet: Donnerstag, 13. August 2009 10:49
An: [email protected]
Betreff: st: AW: Something wrong with the -count- option in the -egen-?
<>
Nothing is wrong, luckily... As -h egen- says about the -count- option: "
creates a constant ... containing the number of nonmissing observations of
exp."
So the surprising result of your last line is easily explained: The
-expression- fed to -count- evaluates to nonmissing -either 0 or 1- for
every one of those five rows of your dataset, and -count- picks up this
number. All it cares about is the fact that the expression is non-missing,
not its content. If you want to -count-, use the standalone -count-...
*************
clear
input x
1
2
3
5
.
end
egen y=count(x)
egen yy=count(x>2)
egen yyy=count(x==2)
//another xmpl,
//20 is not even in the list of values
egen yyyy=count(x==20)
//the following dummy is counted
//it has 5 non-missing entries...
gen byte dummy=x==20
l,noobs
//stand alone -count-
cou if x==2
*************
HTH
Martin
-----Urspr?gliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von [email protected]
Gesendet: Donnerstag, 13. August 2009 10:37
An: statalist
Betreff: st: Something wrong with the -count- option in the -egen-?
I typed the following in the stata and found -count- option didn't work
well.
Anything wrong?
clear
input x
1
2
3
5
.
end
egen y=count(x)
egen yy=count(x>2)
egen yyy=count(x==2)
l,noobs
+------------------+
| x y yy yyy |
|------------------|
| 1 4 5 5 |
| 2 4 5 5 |
| 3 4 5 5 |
| 5 4 5 5 |
| . 4 5 5 |
+------------------+
Thank you for any help!
Best regards,
Rose
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
