Hi. I think that _Error_ retains its value is because of _Error_ is system variable not as the temporary variable which will reset missing at next data step . _Error_ is just like _n_ which is also retain its value at next data step. Ksharp ... View more

Hi. I have try your code.There is no secret about it. If you have sorted the sashelp.class then you will get the output like this: [pre] Name Age Height Weight n1 n2 Alfred 14 69 112.5 10 1 Henry 14 63.5 102.5 11 2 James 12 57.3 83 12 3 Jeffrey 13 62.5 84 13 4 John 12 59 99.5 14 5 Philip 16 72 150 15 6 Robert 12 64.8 128 16 7 Ronald 15 67 133 17 8 Thomas 11 57.5 85 18 9 William 15 66.5 112 19 10 [/pre] That means when the iterative number is ten (i.e. _n_ = 10) ,output an observation,this time n2+1. Ksharp ... View more

Hi. Do you mean in the dataset?But you have two variable id and txt,then 'ABC is my class' will be id or txt? [pre] data temp; input ID $ Type $ Txt & $20.; cards; 1 Class ABC is my class 1 Teacher DEF is my teacher 2 cClass aABC is my class 2 tTeacher dDEF is my teacher ; run; data temp; set temp; _txt=txt; txt=type; if id eq lag(id) then call missing(id); output; txt=_txt; call missing(id); output; drop _txt type; run; [/pre] Ksharp ... View more

Hi. proc format + cnlout option to output the format contents into a dataset. I remember it has mentioned by Art C at before post. Ksharp ... View more

Hi. Art C maybe you misunderstand the op's mean. I think op does not want to process data ,want to Statistical method to process these data. In my opinion, op need SAS/ETS for the sake of time series data. But I am not familiar. It looks like that there is far too much things to learn in SAS.Althought spend your whole life ,it also can not learn it over. Ksharp ... View more

Dear Arther Carpenter: I do not know to say what. Ksharp Message was edited by: Ksharp ... View more

>Bell Curve Hi. Are you to mean Normal Distribution Curve? proc univariate + normal option can do it. Actually It is more suitable that to post it in SAS Graph Forum. There is SAS man named Robert Allison who is expert about Graph. Ksharp Message was edited by: Ksharp ... View more

Hi. If all of the dataset's name are to start with temp ,you can use colon to match all of dataset's name. See the following code,Assuming all the datasets are in the 'c:\'; [pre] libname lib 'c:\'; data all; set lib.temp: ; run; [pre] Ksharp ... View more

Hi. If you know the directory which contains these files. You can use ' filename + pipe 'ls ....' ' which will display the exist files ,then there is not need to use fileexist() function to judge whether the file is existed. and use filevar= just as data _null_; mentioned to input multi-files. Ksharp ... View more

Hi. What do these character variable look like ? Are you using SAS 9.2 with Chinese Version? Ksharp ... View more

Hi. >Another virtue of proc sql is to allow merge on non-matching values That is just because of proc sql 's Cartesian Product. Ksharp ... View more

Hi. Mike Can you explain why it will be appear? It is very interesting. Ksharp ... View more

Hi. Just as Cynthia mentioned before. Let me show your Power of Proc report. [pre] data temp; input ID $ Type $ Txt & $20.; cards; 1 Class ABC is my class 1 Teacher DEF is my teacher 2 cClass aABC is my class 2 tTeacher dDEF is my teacher ; run; ods html file='c:\test.html' style=sasweb ; proc report data=temp nowd out=out; column id type txt; define id /order noprint; define type /order noprint; define txt /order noprint; compute id; if id eq lag(id) then len=4 ; else len=0; endcomp; compute before /style={just=left}; line 'ID ' 'xT'; endcomp; compute after txt/style={just=left}; line id $varying. len type $10.; line txt $20.; endcomp; run; ods html close; [/pre] Ksharp ... View more

Hi. For your situation.I think proc sql is not a good idea. The biggest virtue of proc sql is cartesian product. The data step is much more suitable. One question : you said 'during the 5 days' but your origin data has only 4 days.That is right? [pre] data temp; input ID DATE yymmdd10. VALUE; format date yymmdd10.; datalines; 1 2011/01/01 0 1 2011/01/02 0 1 2011/01/03 1 1 2011/01/04 2 2 2011/01/01 1 2 2011/01/02 2 2 2011/01/03 3 2 2011/01/04 4 ; run; proc sort data=temp; by id date; run; data result; set temp; by id; retain count_three sum_three count_zero .; if first.id then do; count_three=0;sum_three=0;count_zero=0; end; count_three+1; if count_three le 3 then sum_three+value; if value eq 0 then count_zero+1; if last.id then do; if count_zero ge 2 then flag=1; else flag=0; output; end; keep id date sum_three flag; run; proc print;run; [/pre] Ksharp ... View more

Just as Sbb said . Do you try ' input x ?? comma8.' is double ? not single ? Ksharp ... View more