On top of @KevinScott 's and  @SteveDenham 's excellent suggestions, try simplifying the model a bit. You have 132 observations (13 with missing values) and have year, site(year) and block(site year) as random effects. Try removing block(site year) as a random effect and see if that helps. You may not have enough data to get a variance component down to the block level. ... View more

You cannot compute the Mahalanobis distance for your sample data because the correlation/covariance matrix of s1-s8 is fully degenerate. In this case, you cannot invert the correlation or covariance matrix to compute the distances. Computing and utilizing a generalized inverse is not even possible in this case.   If you run PROC CORR, you will see those correlations cannot be computed due to all the values of s1-s8 being 0.   ods output PearsonCorr=sample_corr; proc corr data=sample pearson; var s1-s8; run;     ... View more

The relationship between y and x1 x2 in your data can be captured by using a regression between y and just the quadratic term.  Assuming X2 is X1*X1, try running.   PROC REG DATA=groupD17; MODEL Y = X2; RUN;   And compare the Rsquare with the original model.     Here is an example that will help you understand why a model with only a quadratic term and no linear term can fit well:   data examples(drop=i); do i=1 to 1000; x1=10*ranuni(213)-5; x2=x1*x1; /*y_only_quad is a difference of squares with no linear term y_only_quad=x1^2-4 + N(0,1) error*/ y_only_quad=(x1+2)*(x1-2) + rannor(33); /*y_lin_quad is a quadratic polynomial with a linear term y_lin_quad=x1^2+4x+4 + N(0,1) error*/ y_lin_quad=(x1+2)*(x1+2) + rannor(89); output; end; run; ods graphics; proc sgplot data=examples; scatter x=x1 y=y_only_quad; scatter x=x1 y=y_lin_quad; refline 0 / axis=y; refline 0 / axis=x; run; /*linear term will not be significant assuming alpha level 0.05 rsquare is 0.98*/ proc reg data=examples; model y_only_quad= x1 x2; run; quit; /*remove linear term rsquare is still 0.98*/ proc reg data=examples; model y_only_quad= x2; run; quit; /*run model where the linear term is required the linear term is significant at alpha level 0.05*/ proc reg data=examples; model y_lin_quad= x1 x2; run; quit; ... View more

You are getting the first note because you are not selecting Rownum on the outer Select clause before the nested "except". This is informational because it can indicate that you may expect a result different than you actually get but won't be able to tell why.   The second note is because of the join you used: from names join filter is a Cartesian join forcing every record in Names to be Joined with every record in Filter. If that is not the desired behavior then you use a different join. Again this is partially informational in case you did not intend the Cartesian join as they can be very resource intensive. Two data sets of 1000 records each would require 1,000,000 comparisons for example.   ... View more

Hmmm.  You talk about the objective, but leave it wide open as to what the result should look like.  So here is one way to display the result.   data want; set have; array counts {4} _temporary_ (0 0 0 0); array states {4} CA NY PA FL; do _n_=1 to 4; if states{_n_} then counts{_n_} = 0; else counts{_n_} = counts{_n_} + 1; if counts{_n_} >= 14 then states{_n_} = .F; end; run; This will actually change the data values for CA, NY, PA, and FL.  Any time there are at least 14 periods without a death, the value will change to .F (a special missing value).  You can inspect the resulting data to look for any sets of .F values.  So the output will have the same variables you started with, with .F for the counts when there are at least 14 consecutive no-death days.   Of course, if you have a different outcome in mind, you can always provide details about what it should look like. ... View more

PROC CLUSTER allows you to pass both coordinate data and distance data.   If the data are coordinates, PROC CLUSTER computes Euclidean distances.If you want non-Euclidean distances, You may use the DISTANCE procedure to compute an appropriate distance data set that can then be used as input to PROC CLUSTER. Assuming that you would like to use Euclidean distance then you are not required to do the extra step of converting your data to distances.   If you already have a distance matrix as input then you can use the TYPE= option on a data step to indicate to PROC CLUSTER or KMEANS that the data set is a distance matrix.   https://go.documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=9.4_3.4&docsetId=ledsoptsref&docsetTarget=n1ougt1wsasgafn1xgdfsd8vofoe.htm&locale=en   I hope this information helps. ... View more