Thanks Martin.
Let's take this obs, for example:
idnum wt20 wt30 wt40 wt50 wt60 wt70 weight age_ref weightlst
721 102 147 180 39 30
Using the suggested approach:
gen weight_decades = abs(wt20 - wt30) + ... + abs(wt60 - wt70)
will generate a missing value. But even if it was to work, for this
specific case I would need to add to it the interval abs(147 - 180) because
age_ref > weightlst.
I hope this makes it clearer.
On Tue, 25 Aug 2009 23:02:04 +0200, "Martin Weiss" <[email protected]>
wrote:
>
> <>
>
> I overlooked the missings issue, which -rowtotal- handles more
gracefully,
> sorry about that!
>
>
>
> " In general, if age_ref > weightlst, I need one more interval computed"
>
>
> What should the additional interval contain?
>
>
> HTH
> Martin
>
>
> -----Ursprüngliche Nachricht-----
> Von: [email protected]
> [mailto:[email protected]] Im Auftrag von Martin Weiss
> Gesendet: Dienstag, 25. August 2009 22:45
> An: [email protected]
> Betreff: st: AW: Automating intervals creation
>
>
> <>
>
>
> Your
>
>
> ***
> gen delta1 = abs(wt20 - wt30)
> gen delta2 = abs(wt30 - wt40)
> gen delta3 = abs(wt40 - wt50)
> gen delta4 = abs(wt50 - wt60)
> gen delta5 = abs(wt60 - wt70)
> egen weight_decades = rowtotal(delta1 delta2 delta3 delta4 delta5)
> ***
>
>
> could be easily shortened:
>
>
>
> ******
> clear*
> inp x1 x2 x3
> 1 2 3
> 3 2 5
> 5 9 11
> 2 34 90
> 54 32 9
> end
>
> //easy
> gen weight_decades = abs(x1 - x2)+abs(x2-x3)
>
> /difficult
> gen delta1=abs(x1 - x2)
> gen delta2=abs(x2-x3)
> egen weight_decades2 =rowtotal(delta1 delta2)
>
> //let`s see
> compare weight_decades weight_decades2
> l we*, noo
> ******
>
>
> HTH
> Martin
>
>
> -----Ursprüngliche Nachricht-----
> Von: [email protected]
> [mailto:[email protected]] Im Auftrag von Thomas
Speidel
> Gesendet: Dienstag, 25. August 2009 22:36
> An: [email protected]
> Betreff: st: Automating intervals creation
>
> I have a dataset of this type:
>
+---------------------------------------------------------------------------------------------------+
> idnum wt20 wt30 wt40 wt50 wt60 wt70
> weight age_ref weightlst
>
---------------------------------------------------------------------------------------------------
> 708 61.23497 61.23497 63.50293 61.23497 86.18255 .
90
>
> 60 60
> 713 56.69905 79.37866 81.64663 104.3262 . .
100
>
> 58 50
> 721 102.0583 147.4175 . . . .
180.54
>
> 39 30
>
+---------------------------------------------------------------------------------------------------+
>
> where:
> wt*: weight captured at different ages (10-yr interval)
> weight: weight at reference time
> age_ref: age at reference time
> weightlst: age at last measurement before reference
>
> I am trying to create a measure of weight fluctuation as follows:
>
> -----------------------------
> gen delta1 = abs(wt20 - wt30)
> gen delta2 = abs(wt30 - wt40)
> gen delta3 = abs(wt40 - wt50)
> gen delta4 = abs(wt50 - wt60)
> gen delta5 = abs(wt60 - wt70)
> egen weight_decades = rowtotal(delta1 delta2 delta3 delta4 delta5)
> -----------------------------
>
> I would like to also capture the last interval for some people. For
> example for id==721, I have weight measured at reference (age 39) but the
> last recorded weight is at age 30 (wt30). Thus, the above code will only
> compute 1 interval.
>
> In general, if age_ref > weightlst, I need one more interval computed .
> The preferred way would be to automate my code above to include these
> instances. Thanks for any help.
>
> 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/
>
>
> *
> * 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/
