Hi Anaïs. What you have to do is sort your data set to have the Z observation on the very first observation. Then read the dataset with a DATA step, and [pre] RETAIN Zvalue ; IF _N_=1 THEN Zvalue = value ; [/pre] The RETAIN instruction will ensure that the value is remembered for any further observation read. If there are several values to remember, declare every variable in a single RETAIN, and create them in the IF _N_=1 statement. You can use ARRAYs to simplify syntax, working on week_1, week_2, etc. Regards. Olivier ... View more

Hi all. The solution "transpose-intensive" does not seem to be accurate if your data set is too big (= has too many observations). The first suggestion is more likely to work in any case. The code would look like this... [pre] PROC CONTENTS DATA = dataset1 OUT = variables1 NOPRINT ; RUN ; PROC CONTENTS DATA = dataset2 OUT = variables2 NOPRINT ; RUN ; PROC SQL NOPRINT ; SELECT first.name INTO : commonVariables SEPARATED BY " " FROM variables1 AS first, variables2 AS second WHERE UPCASE(first.name)=UPCASE(second.name) ; QUIT ; /* check the variable names */ %PUT &commonVariables ; DATA dataset1_common ; SET dataset1 (KEEP = &commonVariables) ; RUN ; DATA dataset2_common ; SET dataset2 (KEEP = &commonVariables) ; RUN ; [/pre] Cheers. Olivier ... View more

Hi Bhav. I think the interactive procedures come from the Old Times where it had a real cost to retype parts of programs. The idea behind interactive procs is to easily follow a trial-and-error process ; submit a first draft, see how it turns out, and then a second test, and so on. This is why Datasets, SAS/Graph and some statistical procedures (such as Proc Reg & GLM) are thought this way. For example : [pre] PROC GCHART DATA = sashelp.class ; VBAR age ; RUN ; /* well, as I see the output, it lacks a DISCRETE statement */ VBAR age / DISCRETE ; RUN ; /* I'd better do sub-bars from sex */ VBAR age / DISCRETE SUBGROUP=sex ; RUN ; /* and so on ... */ QUIT ; /* this is the real end of the Gchart procedure */ [/pre] It is quite different for IML and SQL procedures : since you are coding in a different language, the QUIT statement is here to inform SAS that we are coming back to its "usual" programming language. Cheers, Olivier ... View more

This is not really the SQL way of thinking to add grand total lines in a report or a data set. What you can do is same the result of your query in a SAS data set (CREATE TABLE ... AS SELECT etc.) and then display the dataset with a PRINT procedure, using the SUM statement to add the final row : [pre] PROC PRINT DATA = queryResult LABEL NOOBS ; SUM cost ; RUN ; [/pre] Regards. Olivier ... View more

No problem for that. Just change the WHERE part of the query to create (just during the execution of the query) a unique variable that is the concatenation of both dob and name... Just make sure the DOB variable have the same length in both tables, to avoid problems with trailing blanks (otherwise apply TRIM() function to the DOB variable in the query). [pre] PROC SQL ; SELECT COUNT(*) AS howMany FROM a WHERE a.dob!!"/"!!a.name IN (SELECT b.dob!!"/"!!b.name FROM b) /* or TRIM(a.dob)!!"/"!!a.name etc. */ ; QUIT ; [/pre] Olivier ... View more

Hi Maggie. Once again, SQL is your friend for that kind of problem (ID is a variable that is found in both A and B datasets, but the name can vary from a dataset to another, just change in the syntax below) : [pre] PROC SQL ; SELECT COUNT(*) AS howMany FROM a WHERE a.id IN (SELECT b.id FROM b) ; QUIT ; [/pre] Regards. Olivier ... View more

I haven't checked what test KW is exactly, but from what I remember, it is like an analysis of variance on ranks, isn't it ? If so, just compute ranks with PROC RANK and then proceed with PROC GLM. Am I correct ? ... View more

Hi Kwok. The NLEVELS option is new for version 9.0, and I do have information on the option in the SAS Help (press F1 when having your cursor on FREQ in the program). The ODS EXCLUDE ALL prevents SAS from displaying any NLevels table. Since the program is though to be automatic, I have tried to avoid outputs as much as possible -- an old and questionnable habit, I have to admit. If you comment out the EXCLUDE statement, you can also spare the ODS SELECT ALL statement (which turns on output display back to normal). I'm glad the program has help you out. Regards. Olivier ... View more

For the user input, you have several options : 1) if you are using traditionnal SAS environnement, you can check out the %WINDOWS and %DISPLAY syntax to display (ugly) prompt windows and save the values entered in macro variables 2) if you are using SAS Enterprise Guide, you can define parameters (see the OPTIONS menu) that can be used throughout the project, and for which a (very nice) prompt window will be displayed each time the attached task, or query, or code, will be run. ... View more

There is no direct way to limit the number of matches in SQL (or at least, not any I can think of). What you can do, if you have a variable with unique values in data set A, is sort the result of the SQL join by this variable (with an extra ORDER BY clause in the SQL query), and then, if you only want ten matches per value : [pre] DATA myJoinLtd ; SET myJoin ; BY uniqueAvariable ; IF FIRST.uniqueAvariable THEN howManyMatches = 0 ; howManyMatches + 1 ; /* increase the counter */ IF howManyMatches < 11 THEN OUTPUT ; RUN ; ... View more

When I saw Toby mentionning the Freq procedure, I thought it will save time to use the new (version 9) NLEVELS option to automate discarding empty variables. For the empty rows, count non-missing numeric values and test whether the concatenation of all character variables is empty to control deletion. [pre] DATA work.empty_spaces ; INFILE CARDS MISSOVER DSD DLM=" " ; INPUT v1 v2 $ v3 v4 ; CARDS ; 1 1 1 1 2 1 1 3 1 . . ; RUN ; ODS EXCLUDE ALL ; ODS OUTPUT nLevels = work.levels ; PROC FREQ DATa = work.empty_spaces NLEVELS ; TABLE _ALL_ / NOPRINT ; RUN ; ODS SELECT ALL ; PROC SQL NOPRINT ; SELECT tableVar INTO : empty_vars SEPARATED BY " " FROM work.levels WHERE NNonMissLevels = 0 ; QUIT ; DATA work.just_data ; SET work.empty_spaces (DROP = &empty_vars) ; IF N(OF _NUMERIC_) = 0 AND MISSING(COMPRESS(CATS(OF _CHARACTER_))) THEN DELETE ; RUN ; [/pre] Regards. Olivier ... View more

I don't even think that BOTH %sysfunc and %sysevalf are needed : %sysfunc will "host" the computations anyhow. [pre] %LET a = 10 ; %LET b = 100 ; %LET log_a_b = %SYSFUNC(LOG(&a/&b)) ; [/pre] PS : you can even use the %SYSFUNC to round the result, just add a format as a second argument : [pre] %LET log_a_b = %SYSFUNC(LOG(&a/&b), 10.3) ; [/pre] Regards. Olivier ... View more

What you need to separate both lines is to add a "blank" line (an observation with missing values) between the two blocks... And then add the SKIPMISS option in the PLOT statement (causing the lines to pause when meeting a missing value). For example :[pre] DATA work.test ; SET work.test ; BY id ; OUTPUT ; /* save current obs */ IF LAST.id THEN DO ; CALL MISSING(OF l--res1) ; OUTPUT ; /* insert "blank" line */ END ; RUN ; Symbol1 V=CIRCLE I=NONE C=RED; Symbol2 V=NONE I=Join C=BLUE; Symbol3 V=NONE I=Join C=BLUE L=3; PROC GPLOT DATA=test; PLOT L * t = 1 PredL1 * t = 2 CiuL1 * t = 3 CioL1 * t = 3 / OVERLAY SKIPMISS ; TITLE "[1] Anpassung von 2 von-Bertalanffy-Kurven"; RUN; QUIT; [/pre] Regards. Olivier ... View more

When I run your program, I have errors in the Gplot procedures : the variable names are wrong. In the first you, the correct code would be : [pre]PROC GPLOT DATA=test; PLOT L * t = 1 PredL1 * t = 2 CiuL1 * t = 3 CioL1 * t = 3 / OVERLAY; TITLE "[1] Anpassung von 2 von-Bertalanffy-Kurven"; RUN; QUIT; [/pre] (there are "1" missing on three of the variables names) Bad copy & paste, it seems. And same problems with the other two Gplot procs. I don't know if it was the problem you were pointing at (I don't know of Bertalanffy growth curves at all) but I hope it will help... Regards. Olivier ... View more

Check out the PRESS option, in the MODEL statement of the REG procedure. this is related to creating a dataset with coefficients in it (primarly for use in the Score procedure). [pre] PROC REG DATA = sashelp.class OUTEST = work.estimates ; MODEL weight = height age / PRESS ; RUN ; QUIT ; PROC PRINT LABEL NOOBS ; RUN ; [/pre] I don't know if there are equivalents in GLM, LOGISTIC, GENMOD or GLIMMIX procedures (if so, I haven't used them). ... View more