Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: RE: AW: RE: Evaluating a set of conditions


From   Thomas Speidel <thomas@tmbx.com>
To   statalist@hsphsun2.harvard.edu
Subject   RE: st: RE: AW: RE: Evaluating a set of conditions
Date   Wed, 23 Jun 2010 16:58:03 -0600

I have fixed the missing disease below.
I am trying to have as few lines of code as possible and avoid a series of replace. Replacements are easy to follow but they create a hierarchy that is hard to keep track of and error prone. Below I have purposly created a series of replacements. I am looking for the shortest possible way to achieve this that I often see in posts by you and Nick...

======================
input byte(id a b c d e disease)
 1   0   0   1   0   0   0
 2   1   0   1   1   0   1
 3   .   1   1   1   1   .
 4   0   1   1   0   1   0
 5   1   0   0   0   0   0
 6   1   .   1   1   0   1
 7   0   0   0   0   0   0
 8   1   .   .   .   1   .
 9   1   0   0   0   0   0
10   1   .   .   1   1   1
11   1   .   1   0   0   .
12   1   0   1   0   0   0
13   1   0   1   0   0   0
14   .   0   1   0   0   .
15   1   0   1   0   0   0
16   1   0   1   1   0   1
17   0   0   1   0   0   0
18   1   .   .   0   .   .
19   0   .   .   .   1   0
end

egen missing = rowmiss(b c d e)
egen sum = rowtotal(b c d e)
gen byte test = 1 if 	a==1 & 	missing<=1 & 	sum>=2
replace test = 0 if 	a==1 &	missing<=1 &	sum<2
replace test = . if 	a==1 &	missing==1 &	sum==1
replace test = 1 if 	a==1 &	missing==2 &	sum>=2
replace test = . if 	a==1 &  missing==2 &	sum<2
replace test = . if	a==1 & 	missing>=3
replace test = 0 if 	a==0
. assert disease==test

.
======================

Quoting Martin Weiss <martin.weiss1@gmx.de> Wed 23 Jun 16:22:33 2010:


<>

In your -input- call, the "disease" variable seems to be missing.

Anyway, if the only problem is with a certain combination, why not add

***********
replace test=. if a==1 & missing==3
***********


HTH
Martin

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Thomas Speidel
Sent: Donnerstag, 24. Juni 2010 00:09
To: statalist@hsphsun2.harvard.edu
Subject: RE: st: RE: AW: RE: Evaluating a set of conditions

Forcing myself to use -cond-, I get one step closer, yet not quite there
yet:
(in the code below I added one more obs at id==19)

======================
input byte(id a b c d e disease)
  1   0   0   1   0   0
  2   1   0   1   1   0
  3   .   1   1   1   1
  4   0   1   1   0   1
  5   1   0   0   0   0
  6   1   .   1   1   0
  7   0   0   0   0   0
  8   1   .   .   .   1
  9   1   0   0   0   0
10   1   .   .   1   1
11   1   .   1   0   0
12   1   0   1   0   0
13   1   0   1   0   0
14   .   0   1   0   0
15   1   0   1   0   0
16   1   0   1   1   0
17   0   0   1   0   0
18   1   .   .   0   .
19   0   .   .   .   1
end

egen missing = rowmiss(b c d e)
gen byte test =cond(missing(a), ., cond(a==1, cond(missing<=1, (b + c
+ d + e)>=2, cond((b + c + d + e)>=2, 1, .)), 0))

. assert disease==test
3 contradictions in 19 observations
assertion is false
=======================
this seems to fail whenever a==1 & missing==3

Thomas

Quoting Thomas Speidel <thomas@tmbx.com> Wed 23 Jun 13:13:54 2010:

While trying to simplify the problem for the list (my variables are
not actually called a, b, c, etc) I must have inadvertantly
introduced some problems.  Sorry for the confusion.

Nonetheless, the variable called "disease" in the n=18 dataset is
indeed what I am trying to achieve.

Thomas

Quoting Martin Weiss <martin.weiss1@gmx.de> Wed 23 Jun 13:02:40 2010:


<>

Your own code returns "1" for id==11. Have you changed your mind?

***********
clear*

inp byte(id  a b c d e)
1   0   0   1   0   0
2   1   0   1   1   0
3   .   1   1   1   1
4   0   1   1   0   1
5   1   0   0   0   0
6   1   .   1   1   0
7   0   0   0   0   0
8   1   .   .   .   1
9   1   0   0   0   0
10   1   .   .   1   1
11   1   .   1   0   0
12   1   0   1   0   0
13   1   0   1   0   0
14   .   0   1   0   0
15   1   0   1   0   0
16   1   0   1   1   0
17   0   0   1   0   0
18   1   .   .   0   .
end

egen anytwo = rowtotal(a b c d e), missing
egen missing = rowmiss(a b c d e)
replace anytwo = . if (anytwo==0 & missing>=2 & missing<.)
replace anytwo = . if (anytwo==1 & missing==1)
replace anytwo = . if (anytwo==1 & missing==3)
replace anytwo = . if (missing>=4)
gen disease = 1 if (a==1 & anytwo>=2 & anytwo<.)
replace disease = 0 if (a==1 & anytwo<2)
replace disease = 0 if a==0
replace disease =. if a==.


list in 11, noo
***********


HTH
Martin

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Thomas Speidel
Sent: Mittwoch, 23. Juni 2010 17:10
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: RE: AW: RE: Evaluating a set of conditions

Thanks Martin and Nick.  Here is an example where I have added more
missing and manually created "disease" to clarify how the missing
would impact the results:

    id   a   b   c   d   e   disease
     1   0   0   1   0   0         0
     2   1   0   1   1   0         1
     3   .   1   1   1   1         .
     4   0   1   1   0   1         0
     5   1   0   0   0   0         0
     6   1   .   1   1   0         1
     7   0   0   0   0   0         0
     8   1   .   .   .   1         .
     9   1   0   0   0   0         0
    10   1   .   .   1   1         1
    11   1   .   1   0   0         .
    12   1   0   1   0   0         0
    13   1   0   1   0   0         0
    14   .   0   1   0   0         .
    15   1   0   1   0   0         0
    16   1   0   1   1   0         1
    17   0   0   1   0   0         0
    18   1   .   .   0   .         .

Take a look at id==11 for example, where I don't have enough
information to determine disease presence.

Thomas Speidel

Quoting Nick Cox <n.j.cox@durham.ac.uk> Wed 23 Jun 06:59:44 2010:

Yes, if there are missings it's more complicated than my initial answer
could suggest.

(a == 1) & (((b == 1) + (c ==1) + (d == 1) + (e == 1)) >= 2)

would seem to match the possibilities better.

Nick
n.j.cox@durham.ac.uk

Martin Weiss

The result does seem to differ much, though, from the one Thomas
evidently
wants - as expressed by his example:

*************
clear*
set obs 10000
set seed 12345

foreach var of newlist a b c d e{
	gen byte `var'=runiform()<.5
	replace `var'=. if runiform()<.15
}

//NJC
gen disease_true = a & (b + c + d + e >= 2) /*
*/  if !missing(a, b, c, d, e)

//Thomas
egen anytwo = rowtotal(a b c d e), missing
egen missing = rowmiss(a b c d e)
replace anytwo = . if (anytwo==0 & missing>=2 & missing<.)
replace anytwo = . if (anytwo==1 & missing==1)
replace anytwo = . if (anytwo==1 & missing==3)
replace anytwo = . if (missing>=4)
gen disease = 1 if (a==1 & anytwo>=2 & anytwo<.)
replace disease = 0 if (a==1 & anytwo<2)
replace disease = 0 if a==0
replace disease =. if a==.

//Comparison
compare disease_true disease
as  disease_true ==disease
*************

Nick Cox

I think you need to be clear whether missing means true, false or
indeterminate as far as this is concerned.

Setting aside missings, as a, b, c, d, e are Booleans (1 = true, 0 =
false) then

gen disease_true = a & (b + c + d + e >= 2)

is one way to do it. If missings make the problem indeterminate then
tack on

... if !missing(a, b, c, d, e)

Nick
n.j.cox@durham.ac.uk

Thomas Speidel

Following up on my previous post:
http://www.stata.com/statalist/archive/2010-06/msg00984.html
here is an example for something I am trying to achieve in a
nice/efficient/eleganty way.

I have a number of dummies: a, b, c, d, e (missing values do exist)
Disease=true if the following conditions are met:

1) a must be true AND
2) any two of b, c, d, e are true

As I said missing values are crucial, especially when evaluating the
second condition.

My current program works, but I don't think it is efficient and it
probably does things that are unnecessary:

*******************************************
egen anytwo = rowtotal(a b c d e), missing
egen missing = rowmiss(a b c d e)
replace anytwo = . if (anytwo==0 & missing>=2 & missing<.)
replace anytwo = . if (anytwo==1 & missing==1)
replace anytwo = . if (anytwo==1 & missing==3)
replace anytwo = . if (missing>=4)

gen disease = 1 if (a==1 & anytwo>=2 & anytwo<.)
replace disease = 0 if (a==1 & anytwo<2)
replace disease = 0 if a==0
replace disease =. if a==.
*******************************************

I tried to play around with cond, but I found it was making this much
more complicated then it is.  I know I am complicating my life more
than I need to which is why I am looking for alternative solutions.

*
*   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/




--
Thomas Speidel


*
*   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/




--
Thomas Speidel


*
*   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/




--
Thomas Speidel


*
*   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/




--
Thomas Speidel


*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index