On Sun, 18 Mar 2007 20:52:04 -0700, Andrew Hill <hill.andrewd@GMAIL.COM> wrote:

>Sorry Howard. > >I am not sure why that happend. > >Here is the text again: > > Hi Howard, > >Thanks for your response. > >The reason I don't have a monthly interval is because the trees in the >stands are the analytic unit. I will be using the reduced weather data to >model diameter changes in the trees. > >The variable PRC_50_79_jan is the average precipitation that fell in January >from 1950 to 1979,

That's 30 years, including both endpoints. Yet elsewhere you say that the "measurement intervals" range from 1 to 12 years.

Also, aren't there geographic coordinates for the locations where the measurements were taken?

>not important for solving the coding part of the problem, >but it is the variable name. > >Your suggestion of normalizing is a good one. In fact I have several >different data sets that are based on the measurement interval for the >trees. > >The ideal solution is one the keeps the ID, measurement interval, and first >measurement year, then simply has the variables that I need listed as >Y1_Jan_precip Y2_jan_precip and so on. > >From their I can manipulate them as I need to come up with a solution to my >modeling problem. > > > >On 3/17/07, Andrew Hill <hill.andrewd@gmail.com> wrote: >> >> Dear SAS list, >> >> >> >> I am in the middle of a problem that I could solve via lots of repetitive >> coding, but I think there may be an easier way that one of you may know. >> >> >> >> My problem is this: >> >> >> >> I have a data set with about 3000 variables. (Monthly weather data on a >> 2X2 Km grid for WA, if you are interested). >> >> >> >> Information from this dataset will be integrated with some data from the >> FS about specific stands of trees. The problem that I am facing is that the >> measurement intervals range from 1 year to 12 years in length. Additionally, >> the first measurement year ranges from 1988 to 2001. >> >> >> >> So what I want to end up with is a data set that looks like this: >> >> >> >> ID Init_meas Measurement interval, V1, v2, v3��V36. >> >> >> >> Right now I am looking at writing code that does this with a brute force: >> >> >> >> DATA CLIANL; >> >> SET CLIMIT; >> >> IF INIT_MEAS = '1988' THEN V1 = PRC_50_79_JAN; >> >> IF INIT_MEAS = '1988' THEN V2 = PRC_50_79_FEB; >> >> . >> >> . >> >> . >> >> IF INIT_MEAS = '1993' THEN V1 = PRC_55_84_JAN; >> IF INIT_MEAS = '1993' THEN V1 = PRC_55_84_FEB; >> . >> . >> . >> LOTS MORE OF THE SAME; >> RUN; >> >> >> >> There has to be an easier way to do this. Something that relies on the >> recursion in the process, but I am not sure how to do it. >> >> >> >> Any help is appreciated. >> >> >> >> Thanks, >> >> >> >> Andrew >>