How were these datasets created? Did you create them? If so, the best solution is to back up a step and avoid the issue in the first place. ... View more

You don't show any numbers in your expected output so it is impossible to tell what "percentage sum" you want to calculate.   Also it looks like you have unique combinations for A, B and C so the counts of any frequency tables are going to be 1 in every combination so I am not sure what you intended to calculate. ... View more

@Gick wrote: In fact when there are more than 10 tables, the code does not take into account non-empty ones. @Gick That's not the solution and what you observe has NOTHING to do with the number of tables. ... View more

And you don't need to re-count the number of names.  PROC SQL already counted them for you. so just remember that number. proc sql; select nliteral(name) into :char_vars separated by '|' from dictionary.columns where libname='WORK' and memname="MATABLE" and type='char' ; %let nvars=&sqlobs; quit; %do i=1 %to &nvars; ... ... View more

What do you mean by "create a macro program" ? Is Cal itself a macro  program, starting with %macro <name>(...); and ending with %mend ? Or does it generate a macro program? In first case you can just %include <path>/Cal.sas at the top of any program you want to use Cal macro program. The mcaro will be compiled at submission and used when called. In second case you need to keep the generated macro progam as a <filename>.sas in a folder then use it as described above. ... View more

/*Another code*/ %macro remplacer_points(table_in, table_out); proc sql noprint; select name into :varlist separated by ' ' from dictionary.columns where libname = upcase(libname) and memname = upcase("&table_in"); quit; data &table_out; set &table_in; %let num_vars = %sysfunc(countw(&varlist)); %do i = 1 %to &num_vars; /* Obtenir le nom de la variable */ %let var_name = %scan(&varlist, &i); %put &var_name.; %let new_var_name = %sysfunc(tranwrd(&var_name, '.', '_')); %put &new_var_name.; rename &var_name = &new_var_name; %end; run; %mend; Always the same ... View more

What is the error message you receive?   You could have missed a semicolon or compiled the macro wrong....in fact your code doesn't show the macro compilation step.  Which is why we need to see the log.   @Gick wrote: Hi, I get an error message that I don't understand when I use the SAS CALMAR macro to calculate the weights. I read the documentation and apply it but it doesn't work. Even the example given in the documentation does not work for us. Here is my data and the Squid macro. Can anyone help me or someone who has already used this SAS Calmar macro can explain to me what is wrong in my code or if I forgot a line of code before running the macro SAS CALMAR. Thanks in advance. Gick data BDD_recod; input Nom $ sexe $ X $ Pond ; datalines; Jean Masculin 0-4ans 10 Marie Feminin 5-14ans 10 Pierre Masculin 15-24ans 10 J Feminin 0-4ans 10 M Feminin 5-14ans 10 Pe Masculin 0-4ans 10 Je Masculin 15-24ans 10 Ma Masculin 0-4ans 10 Pire Feminin 25-30ans 10 Mfir Feminin 35-40ans 10 Pell Masculin 40-45ans 10 Jeurs Masculin 15-24ans 10 Mam Masculin 0-4ans 10 Piress Feminin 15-24ans 10 ; data marges_cibles; input VAR $ N MAR1-MAR6 ; Cards; sexe 2 67866 80149 . . . . X 6 33065.5 39023 35281 21382 11561 7702 ; %CALMAR(DATA=BDD_recod, POIDS=Pond, IDENT=Nom, DATAMAR=marges_cibles, M=2,EDITPOI=OUI,OBSELI=OUI,DATAPOI=SORTIE,POIDSFIN=PONDFIN,LABELPOI=pondération raking ratio)     ... View more

@Gick wrote: this is how my starting table looks   Thanks for your help. Gick, You keep using that word.  Since we cannot program from a picture or from a table in a report let's assume you are using "table" to mean a dataset.  So you want to start with dataset in this picture and produce a dataset that would look like the other picture.   You just need use RENAME= and FIRSTOBS= dataset options. To generate the old=new name pairs from the first dataset you can use PROC TRANSPOSE and PROC SQL to create a macro variable. proc transpose data=HAVE (firstobs=3 obs=3) out=names; var _all_; run; proc sql noprint; select catx('=',nliteral(_name_),nliteral(col1)) into :renames separated by ' ' from names ; quit; Now you can use that macro variable in a step that skips over the first three observations of the original dataset. data want; set have(firstobs=4 rename=(&renames)); run; ... View more

Follow the link I provided in my first post to PROC GCHART and the PIE statement. ... View more

Read "Construct a stacked bar chart in SAS where each bar equals 100%"   If you don't need the labels for each segment, PROC FREQ can create the entire plot. FOr more control, use PROC SGPLOT, as shown: proc transpose data=piechart Name=Period out=temp(rename=(col1=Count)); by secteur; var Annee_2017 Annee_2018 Annee_2019; run; /* stacked, but no bar labels */ proc freq data=temp; tables Period*secteur / plots=freqplot(twoway=stacked scale=grouppct); weight Count; run;   ... View more

The idea of arranging data in this fashion seems to be a poor idea. Almost all SAS PROCs expect data to be arranged vertically, not horizontally. Yes, you can do additional programming to make it work horizontally, but its much much easier with vertical data sets, which is why I keep stealing Maxim 19   Maxim 19 Long beats wide. (Don't keep data in structure) In the world of spreadsheets, people tend to line up data side-by-side, and put data items (dates, categories, …) into column headers. This runs counter to all the methods available in SAS for group processing, and makes programming difficult, as one has variable column names and has to resort to creating dynamic code (with macros and/or call execute) where such is not necessary at all if categories were represented in their own column and data aligned vertically. There are times where a wide format is needed, eg when preparing data for regression analysis. But for the processing and storing of data, long formats are always to be preferred. Dynamic variable names force unnecessary dynamic code.     data have; input name $ time value; datalines; John 1 0 John 2 1 John 3 3 ... David 16 32 ;   From here, the programming is a lot easier than working with variable names like '4_1'n, using PROCs like PROC SUMMARY, and maybe even SQL works a lot easier. And while you have marked an answer as "Correct", I think the whole approach is a poor way to do this. ... View more

I cannot figure out what you are trying to do. You are running a DO loop based on the number of words in the string using COUNTW() with the default set of delimiters:   If your computer uses ASCII characters, the default delimiters are as follows: blank ! $ % & ( ) * + , - . / ; < ^ | But later you are calling the %SCAN() with just two characters as delimiter, space and double quote.   Also why are you having the macro generate SAS statements?  Are you calling this macro in the middle of a DATA step or some other context where an IF statement makes sense?   And the IF statement you are generating also looks a little strange. if index(%scan(&list_debug.,&d.," "),'&apos;') > 0 then %scan(&list_debug.,&d.) = tranwrd(%scan(&list_debug.,&d.),'&apos;'," ") ; It seems to be assuming that the next word in the list is a valid SAS variable name.  (but notice that the variable using the IF condition is potentially different than the variable used in the generated assginment statement because using different delimiters in the %SCAN() calls). So I think you are trying to generate a statement like this: if index(A,'&apos;')>0 then A=tranwrd(A,'&apos;',' '); Which you does not need to IF, if the string is not found then tranwrd() already does nothing. A=tranwrd(A,'&apos;',' '); If also want to convert actual single quotes instead of just the HTML command to generate a single quote you can use: A=tranwrd(tranwrd(A,'&apos;',' '),"'",' '); ... View more