This is kind of basic, but I'm having trouble creating a new variable with three different conditions, and can find very little documentation. For example:   data new; set old;   if var1='T' and var2='T' and var3='T' then newvar='cat1'; if var1='T' and var2='F' and var3='F' then newvar='cat2'; if var1='T' and var2='T' and var3='F' then newvar='cat3'; if var1='F' and var2='F' and var3='F' then newvar='cat4';   run;   It seems straightforward, but the output will only incorrectly code one or two categories. Originally I was using if-then-else statements (which also didn't work), but I read that when coding for multiple conditions, each statement should only start with if. I've also tried putting brackets around the categories to delineate them. I'm working in linux, if that makes a difference.    Thanks! ... View more

I'm trying to figure out how SAS calculates the per level estimates using contrast statements, as well as the predicted survival per person in a survival model with an interaction coefficient.    For example, I'm trying to predict an outcome from bmi, a binary categorical variable (1 vs. 2), a binary covariate (0 vs. 1), and an interaction term (bmi_lvl=bmi*lvl). Here's the data:     data temp; input ID foltime bmi cov lvl bmi_lvl outcome; datalines; 1 1.2 20 0 2 40 1 2 0.5 23 1 2 46 1 3 4.5 30 0 1 30 0 4 3.6 18 0 2 36 0 5 1.0 25 0 1 25 1 6 1.2 22 0 1 22 1 7 4.0 26 1 2 52 0 8 3.3 26 0 1 26 1 9 0.8 24 1 2 48 0 10 1.2 32 0 2 64 0 11 3.0 22 1 2 44 1 12 3.3 23 1 2 46 1 13 1.0 30 0 1 30 0 14 1.8 36 0 2 72 1 15 0.9 25 1 2 50 1 ; run;     I plug it into the model:   proc phreg data=temp outest=beta; class cov (ref='0'); model foltime*outcome(0)=bmi lvl cov bmi_lvl/ties=efron RL; *output statement provides the per person estimated survival; output out=pred survival=survpred; *contrast statement provides hazard estimates at specific BMI and level, at reference category for covariate; *contrasting hazard at BMI 24 for levels 1 and 2; contrast "lvl 1 at bmi 24" lvl 1 bmi 0 cov 0 bmi_lvl 24/estimate=exp e; contrast "lvl 2 at bmi 24" lvl 2 bmi 0 cov 0 bmi_lvl 48/estimate=exp e; *contrasting hazard at BMI 28 for levels 1 and 2; contrast "lvl 1 at BMI 28" lvl 1 bmi 0 cov 0 bmi_lvl 28/estimate=exp e; contrast "lvl 2 at BMI 28" lvl 2 bmi 0 cov 0 bmi_lvl 56/estimate=exp e; run;     I set BMI to 0 in the contrast statement based on example 2, page 10 of this paper, https://support.sas.com/resources/papers/proceedings10/253-2010.pdf, however, I'm not sure why that is.    The coefficients are -0.8357 for BMI, -11.398 for lvl, 0.4375 for cov, and 0.4345 for bmi_lvl. Attached are the results for the contrast statements, as well as the per person estimated survival.    How does SAS calculate these?   Thanks!           ... View more

Does anyone know how SAS calculates Schoenfeld residuals in survival analysis? Are they scaled? My understanding is that it's the value of a covariate for a given individual subtracted by the weighted average of that covariate among individuals who failed (i.e. got the outcome of interest) at time T. I would think that in order to get the weighted average, you would need to multiply the model coefficient times the individual covariate value, and then add that to the covariate to get an expected value, which you would then take the average of. The average expected value would be among those who did and did not get the outcome of interest at time T. However, when I tried calculating this by hand, it didn’t match with the results. I’m also unsure as to how SAS calculates a residual for binary or categorical variables.   Any insight would be appreciated. Thanks in advance! ... View more

Hi,   I have a question about proc sgplot. I created a graph which has the probability of an outcome by an x value, and then I added another x series to show the point where the maximum probability is attained. Here is my code:     proc sgplot data=ex dattrmap=attr1; series x=x_value y=probability/ group=Group linattrs = (thickness=3px) attrid=Group; series x=max y=probability/group=Group Datalabel=Max markers markerattrs=(symbol=circlefilled size=10px) attrid=Group; yaxis integer values=(0 to 0.11 by .01); xaxis integer values=(1 2 3 4 5 5.10) offsetmax=.01; run;   The problem is that I cosmetically need the x axis to show only the values 1 to 5, but then for it in reality to go slightly over 5 to show the point where the probability reaches the maximum (5.10). The code above gives me the x-axis values 1 3 and 5, for some reason. I would like it to say 1 through 5, but still show the point 5.10 on the graph. Does anyone know how to do this?   Thanks in advance!   ... View more

Hi,   My question requires IT, rather than code support. I'm using SAS Enterprise Guide version 7.12 HF4. There is an option to create projects, and within each project there is a column on the side bar labeled Project Tree with multiple SAS files.   I was in the process of creating a new project, and trying to copy and paste code from a previous project to the new project. When I saved all, the files in my original project were completely deleted and restored to older versions.When I go to the folder directory corresponding to my libname statement, the newest file is from 2017.    Obviously I did something, but I don't know what. I am also confused as to what the "Process Flow" is for. When I click on the tables, it just tells me that the files don't exist, or that there's a data operation error. I don't know if I clicked on an older version of the process flow, and then accidentally saved. There are also a few files under 'Programs' which give me errors that "The file associated with this view could not be located". I attached a picture of what the project tree looks like.    Has this ever happened to anyone, and if so, have they been able to retrieve their files?   Thanks in advance,   Emily    ... View more

Hi,   I am trying to select only observations that fall on the tenth, as opposed to hundredth decimal place (e.g. 1.10, 1.20, 1.30 vs. 1.11, 1.12, 1.13). The variable goes from one to five.    I tried:   data want; set have; do i=1 to 5 by .1; where var=i; end; run;   I receive an error that i is not on file in my data set. Does anyone know why/a different approach?    Thanks in advance,   Emily  ... View more

Hi,   I am trying to add 16 rows to a table, where variable x1 goes from 1 to 5 by .25 increments. x2 is the square of x1, and the remaining variables are the same (either 0 or the mean).   For example,   ID            x1 x2 x3 x4 x5 x6 x7 .               .   .   .    .   .   .    . .               .   .   .    .   .   .    . .               .   .   .    .   .   .    . Dum1      1  1  0   0  0  0  27.9 Dum125 1.25 1.56 0 0 0 0 27.9 Dum150 1.50 2.25 0 0 0 0 27.9 Dum175 1.75 3.06 0 0 0 0 27.9 ...   I originally used the input statement to create the data (above), and then appended it my original dataset. Does anyone know a better way?   Thanks in advance,   Emily ... View more