1. Can I code at SAS OnDemand to upload my local files? Or do I always need to use my mouse to upload something? For example, the following code enables SAS OnDemand to read FRED data without clicking a mouse. data will5000indfc; infile "https://fred.stlouisfed.org/graph/fredgraph.csv?id=WILL5000INDFC" url firstobs=2 dsd truncover; input date yymmdd10. +1 will5000indfc; run; Or for example, I can code at SAS OnDemand to directly download whatever data on the Internet as follows. data _null_; infile "https://www.mathematik.uni-wuerzburg.de/fileadmin/10040800/user_upload/time_series/data_sets/iq.sas7bdat" url recfm=f; input; file "%sysfunc(getoption(work))/iq.sas7bdat" recfm=n; put _infile_; run; What if the files I want to analyze are in my C drive? I wonder whether I can automate the processes using code at SAS OnDemand. 2. Can I code at SAS OnDemand to download some results? Suppose I analyzed something at SAS OnDemand and want to download the results as follows. data will5000indfc; infile "https://fred.stlouisfed.org/graph/fredgraph.csv?id=WILL5000INDFC" url firstobs=2 dsd truncover; input date yymmdd10. +1 will5000indfc; run; data tedrate; infile "https://fred.stlouisfed.org/graph/fredgraph.csv?id=TEDRATE" url firstobs=2 dsd truncover; input date yymmdd10. +1 tedrate; run; data residual; merge will5000indfc tedrate; by date; run; proc reg noprint; model will5000indfc=tedrate; output residual=residual out=residual; quit; Can I code at SAS OnDemand to download the resultant  residual data set? Or suppose I exported the data set to a csv file as follows. proc export replace file="/home/junyong/residual.csv"; run; Do I need to use my mouse to download the residual.csv ? Or can I code at SAS OnDemand to download this file automatically to, for example, my desktop? 3. Is there a URL for the residual.csv such as, for example, https://welcome.oda.sas.com/junyong/residual.csv ? If this is possible, then I will be able to use the second code above to download the results locally. 4. Can one SAS OnDemand user access to the data of another SAS OnDemand user? I wonder whether multiple SAS OnDemand users can share their uploaded data interactively. ... View more

I use class in proc glm to run a one-way fixed-effect model as follows. data have; do i=1 to 3; do j=1 to 3; do k=1 to 10; x=rannor(1); y=x+rannor(1); output; end; end; end; run; proc glm data=have; class i; model y=x i/noint solution; ods output parameterestimates=oneway_class; quit; Or sometimes I use absorb because it is faster than class . proc glm data=have; absorb i; model y=x/noint solution; ods output parameterestimates=oneway_absorb; quit; And I found that absorb , unlike class , cannot estimate a two-way fixed-effect model in this way. /*THE FOLLOWING ESTIMATES TWO-WAY FIXED-EFFECT MODELS CORRECTLY*/ proc glm data=have; class i j; model y=x i j/noint solution; ods output parameterestimates=twoway_class; quit; /*THE FOLLOWING DOES NOT ESTIMATE TWO-WAY FIXED-EFFECT MODELS*/ proc glm data=have; absorb i j; model y=x/noint solution; ods output parameterestimates=twoway_absorb; quit; Though the first glm with class is correct, it becomes too slow because the real data have so many observations. How can I estimate a two-way fixed-effect model with absorb in glm instead? ... View more

I saw Proc Reg several regressions with missing values that talks about multiple model s with some missing y s. In short, proc reg with multiple model s excludes observations with either y1 or y2 missing from all of its model s. My case is opposite—my proc reg with multiple model s has some x observations missing as follows. data have; do i=1 to 5000; x1=rannor(1); x2=rannor(1); x3=rannor(1); y=x1+x2+x3+rannor(1); if ranbin(1,1,0.01) then x1=.; if ranbin(1,1,0.01) then x2=.; if ranbin(1,1,0.01) then x3=.; output; end; run; And I want to make each model in the following proc reg use all available observations. proc reg noprint outest=want; model y=x1; model y=x2; model y=x3; model y=x1 x2; model y=x1 x3; model y=x2 x3; model y=x1 x2 x3/edf; quit; I cannot apply the method above here because my model has different x s rather than y s. Is separating proc reg s the only solution? ... View more

I have the following panel data of 500 individuals for 60 months. data have; do i=1 to 500; do j="1jan2016"d to "31dec2020"d; x=rannor(1); if j=intnx("month",j,0,"end") then output; end; end; run; And suppose I want to add lag(x) with SQL as follows. proc sql; create table want as select a.*,b.x as x1 from have a join have b on a.i=b.i and intnx("month",a.j,0)=intnx("month",b.j,1); quit; And I found that altering intnx("month",a.j,0)=intnx("month",b.j,1) by intck("month",a.j,b.j)=-1 makes the proc sql slower—why is this the case? ... View more

I have 5000 daily time series ( var0001-var5000 ) from 01/01/2019 to 12/31/2019 and want to compute their rolling skewness each month. This post introduces proc summary for the rolling skewness. data have(drop=i); format t yymmddn6.; array var(5000) var0001-var5000; do t="1jan2019"d to "31dec2019"d; do i=1 to 5000; var(i)=rannor(1); end; output; end; run; proc summary nway; class t; format t yymmn6.; var var0001-var5000; output out=want skew=; run; So I can quickly compute the 5000 skewness in January, February, and so on. Now suppose I want to compute each monthly skewness using two—or N more generally—consecutive months rather than just one month—for example, the February skewness using January and February data, the March skewness using February and March data, and so on. Can proc summary do this? ... View more

I am using npar1way to get Kolmogorov–Smirnov statistics and want to get the empirical distribution functions as well. However, I am already using ods listing close and ods results=off to suppress both output and results as follows. data have; do i=1 to 5; do j=1,2; do k=1 to 20; x=rannor(1); output; end; end; end; run; ods listing close; ods results=off; proc npar1way edf d plots=edf; by i; class j; var x; ods output ks2stats=out; run; ods results=on; ods listing; I want to use ods listing gpath="!userprofile\desktop\" and ods graphics/imagename="edf" to draw the five (for i=1 to 5 ) pictures—with output and results still suppressed. Where should I locate the ods listing gpath then? Thanks. ... View more