proc sql; select * from T2 natural inner join ( select A,C,D,F from T2 except select A,C,D,F from T1 ) ; quit; ... View more

@MILKYLOVE wrote:   How do I point to a file repository if I don't have one? You can't.   Your logic is flawed somewhere. The code requires reading something from the file identified as FIC, and yet you don't have such a file. You tell me how to fix this in words (not in SAS code), because I don't know what you are trying to do. ... View more

@MILKYLOVE wrote: Hello   For some reasons I can't seem to do this. How do I convert a date DATE = '20231002' to DATE_EXTRACTION  = '02OCT2023' in SAS please?    I have tried :   DATE_EXTRACTION = put(datepart(DATE), date9.);   But I get 22AUG1960 instead of 02OCT2023     Dates should never be character strings, they should be numbers. Numbers that represent the number of days since 01JAN1960.   Had you written this   DATE = '02OCT2023'd; format date date9.;   you would have the correct date ... this is the way to translate dates that you understand to numerical SAS dates. Also please note that this will not work:   DATE = '20231002'd;   Lastly, when you have actual SAS dates, you do NOT need the DATEPART function; the only time you need the DATEPART function is when the value is SAS date/time values. But your value is a SAS date, not a date/time value, and DATEPART will not do meaningful things to date values.     ... View more

ONE case statement assigns ONE value to ONE variable. So your code is apparently trying to violate that. CASE WHEN CD_ROLE = ' ' THEN r.SOURCE ELSE CD_ROLE END AS CD_ROLE length = 20 and a second case for the CASE WHEN CD_ROLE = ' ' THEN r.NU_ROLE else ' ' /* guessing as you did not provide an ELSE rule for NU_CLI*/ END AS NU_CLI length = 20   ... View more

Thank you so much!!    This worked ... View more

Thank you !! It works 🙂 ... View more

@MILKYLOVE wrote: The problem with this solution is that not all the variables of each array list are of the same data type.  Therefore I am getting errors.   The program also does not recognise '-' That rule doesn't make sense for a numeric variable. In SAS, the variable would not be able to have - characters in the first place. If you need to convert types that's a more complex rule.    Cleaning data is always a somewhat tedious process. You have two options, for each rule loop through the variables, or for each variable loop through the applicable rules. It's easier to do the first one in SAS via arrays.      ... View more

I already showed you how to generate the split files with numeric suffixes using a different method.   If you want to rename the existing files then first get the list of files into data.  Say you have the list in a dataset named FILES and the filename is in a variable named FILENAME.  Then you could use code like this to generate NEW_NAME variable. proc sql; create table rename_files select substr(filename,1,length(filename)-6) as basename , count(*) as n_files , filename group by 1 having count(*) > 1 ; quit; data rename_files; do suffix=1 by 1 until(last.basename); set rename_files; by basename; new_name = cats(basename,putn(suffix,cats(Z,length(cats(n_files)),'.'),'.csv');       output;    end; run; You could then use that list to generate code to move, rename or copy the files. ... View more

Please pay attention to the hint from @andreas_lds , you don't need to specifically name each .csv file in your code. You don't even have to read the list of names in the folder. ... View more

The data you have posted is a perfect demonstration why relying on proc import is a bad idea. Instead of reading the first variable as char, the contents of the variable causes sas to set the wrong type. So: don't use proc import write a data step yourself.    Untested: data work.scores; infile "C:/mydir/scores_*.csv" delimiter= "|" firstobs= 2 missover; length number $14 dab rc construction flotte snt_col pre_col $ 1 activities $ 200 ; input number -- activities; run; If you want to see the real contents of a csv file, open it using a text editor, not Excel. ... View more

Hello @MILKYLOVE,   Add the appropriate ORDER BY clause to the PROC SQL step creating work.t2a_extract_email: order by SOURCE, NU_ODS_PER_ROLE; as you did in your first PROC SQL step. Or, shorter and similar to your second PROC SQL step: order by 1, 2; ... View more

Suggest, if you are only interested in a few variables that you add a VAR statement to the Proc Compare code. The VAR statement will restrict the comparisons to only the variables on the statement. So supply the date variables of interest.   With Proc Compare you also want to sort the data sets in some fashion that the order makes sense for comparing records. I am not sure that dates alone are appropriate if you want to see if the dates associated with some characteristic such as company, person, instrument or other grouping. ... View more

If your system is using ASCII characters then Sepchar=3B should work. ... View more

KISS (Keep It Simple, Stupid) principle: since your path names have no blanks, no quotes are needed around them. Macro triggers will not be resolved inside single quotes. And the X statement is a very crude tool, not giving you any more information than the return code in &SYSRC.   In light of this, do %let RacineRun = /sasprep/fit/fit_1/run/azi/d2d/w6; %let RacineBloctel = particuliers/bloctel/fichiers; /* you use a slash later when you call the macro variable */ %let REP_BLOCTEL_RETOUR = /sasprep/prep/cft/w6/BLOCTEL/ENTRANT; data _null_; infile "cp &rep_bloctel_retour./* &RacineRun./&RacineBloctel./archive 2>&1" pipe; input; put _infile_; run; The INFILE PIPE method shown here retrieves all responses from the external command to the SAS log. The 2>&1 at the end of the command reroutes stderr to stdout, so the pipe catches "normal" and error messages. If you still get issues, post the complete log from the above code. ... View more