Hello,   You can use the vformat function to check the formats of the elements in an array. Also, you cannot change the format of a dataset column on the fly. You have to create a new column and drop the old one.   data mpart; format BDAY REcDate ReTDate datetime.; BDAY=datetime(); RecDate=datetime(); RetDate=datetime(); x=1; run; data _NULL_; set work.mpart; array fixed _NUMERIC_; call execute('data want; set work.mpart;'); do over fixed; if vformat(fixed)=:"DATETIME" then do; varname=vname(fixed); call execute(cat('drop ',strip(varname),';')); call execute(cat('format N_',strip(varname),' date9.;')); call execute(cats('N_',varname,'=datepart(',varname,');')); end; end; call execute('run;'); stop; run; ... View more

Hello,   This is what the proc freq documentation says about the order=FORMATTED option   https://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_freq_sect006.htm FORMATTED orders values by their formatted values (in ascending order). This order is dependent on the operating environment.   So the alphabetical order of formatted values will be used which is not what you want. What you want, as i understand it, is the order in which formatted values have been defined in the proc format.   The following program exports format definitions in a SAS dataset and uses this dataset to impose a custom sort order on the input dataset. Proc freq is then used with the order=data option (i used the program you gave in your other post on the same subject).   proc format; value $Fm1t ( notsorted) 'Male'='M' 'Female'='F'; value $Fm2t ( notsorted) 'Desirable'='Desirable' 'Borderline'='Borderline' 'High'='High'; value $Fm3t ( notsorted) 'Underweight'='Underweight' 'Normal'='Normal weight' 'Overweight'='Overweight'; value $Fm4t ( notsorted) 'Non-smoker'='Non-smoker' 'Light (1-5)'='Light Smoker (1-5)' 'Moderate (6-15)'='Moderate Smoker(6-15)' 'Heavy (16-25)'='Heavy Smoker(16-25)' 'Very Heavy (> 25)'='Very Heavy Smoker(> 25)'; value $Fm5t ( notsorted) 'Optimal'='Optimal Blood Pressure' 'Normal'='Normal Blood Pressure' 'High'='High Blood Pressure'; select $Fm1t $Fm2t $Fm3t $Fm4t $Fm5t; Run; Data heart; set sashelp.heart; Format sex $Fm1t. Chol_Status $Fm2t. Weight_Status $Fm3t. Smoking_Status $Fm4t. BP_Status $Fm5t.; Run; %macro mmacro1(pred,i); proc format library=work.formats fmtlib out=desc; select $Fm&i.t; run; data sort_ds; set desc; value=start; n=_N_; keep value n; run; proc sql; CREATE TABLE heart_s AS SELECT a.* FROM heart a LEFT JOIN sort_ds b ON b.value=a.&pred. ORDER BY b.n; quit; proc freq data=heart_s order=data; tables &pred*Status/chisq sparse outpct out=outfreq&i ; output out=stats&i chisq; format &pred $Fm&i.t.; run; %mend; %mmacro1(sex,1); %mmacro1(Chol_Status,2); %mmacro1(Weight_Status,3); %mmacro1(Smoking_Status,4); %mmacro1(BP_Status,5);   ... View more

Hello, SAS doesn't allow to create a dataset with cards or datalines in a macro. See "Details" section in http://support.sas.com/documentation/cdl/en/mcrolref/61885/HTML/default/viewer.htm#macro-stmt.htm ... View more

Hello,   What do you need that for ? If you need more that three ampersands to resolve a macrovariable there is probably a design problem in your code.   %let TRUE=GE; %let GE=AND; %let AND=LE; %let LE=TRUE; %macro MinAmpersand(var); data _NULL_; length value $4. amps $20.; var="&var."; amps="&"; do while (var ne value and MAX_LOOP<20); MAX_LOOP+1; value=resolve(cats(amps,"TRUE")); amps=cats(amps,"&"); end; nbamps=length(amps)-1; put nbamps " ampersands necessary for macrovariable &var."; run; %mend; %MinAmpersand(TRUE); %MinAmpersand(GE); %MinAmpersand(AND); %MinAmpersand(LE); ... View more

Hello,   data pQ3; input id Country $ PriceQ3; cards; 44 France 3 45 Allemagne 5 46 Belgique 6 47 Espagne 7 ; run; data pQ4; input id Country $ PriceQ4; cards; 44 France 45 45 Allemagne 21 46 Belgique 6 47 Espagne 7 48 Russie 34 ; run; proc sql; CREATE TABLE P_hist AS SELECT COALESCE(a.id,b.id) AS id, COALESCE(a.Country, b.Country) AS country, a.PriceQ3, b.PriceQ4 FROM pQ3 a FULL JOIN pQ4 b ON a.id=b.id; quit;   Edit : added names for the calculated columns (id, Country) ... View more

Hello, Can you post the complete log for an execution of your program ? ... View more

Hello,   with call execute in a data step :   %let table=sashelp.class; %let column=age; data _NULL_; input age; call execute(cats('data out_',&column.,'; set &table.; where &column.=',&column.,'; if _N_ le 3; run;')); cards; 12 13 14 15 16 ; run; ... View more

Hello,   Can you show your code ?   If i'm not mistaken, SAS University edition runs on a linux VM. So if you run it on a windows machine, your file has to be on a shared directory mounted by the virtual machine and the path specified accordingly. ... View more