Hi all,   I currently process strings that contain comma seperated substrings.   The Example below demonstrates that I can succesfully access each of these substrings using the SCAN function... ...unless there is a substring (here, it is the 3rd one) that was enclosed in quotations marks because it contains the delimiter, i.e. the comma. In this case, the substring is only partially read until the comma occurs. Usually, it is suggested to use the q modifier to account for this situation. Unfortunately, the q modifier (just remove the comment symbols from the code below) does not work for me because of the unmatched quotation marks that can occur in the strings.   Is there an option that I'm missing or an elegant alternative to account for both: (1) substrings containing the delimiter enclosed in quotation marks and (2) unmatched quotation marks?   The wanted outcome is: 1: Barcelona 2: Jack's Place 3: Blue Sea, LA   As always, any help is appreciated!     data _null_; have = "Barcelona,Jack's Place,""Blue Sea, LA"""; put have; do i=1 to 3; want = scan( have, i, ',', /*'q'*/ ); put i ":" want; end; run;     ... View more

Hello everyone, When I... * have size values ranging from, let's say, 4 (minimum) to 20 (maximum) * and then define a radius range from 4pt (BRADIUSMIN=4pt) to 20 pt (BRADIUSMAX=20pt) ...I always thought that any size value x between 4 and 20 would result in a bubble of radius x pt. However, the example below shows that this doesn't seem to be the case: The size values for the red bubbles range from 4 to 20. Therefore, BRADIUSMIN=4pt BRADIUSMAX=20pt. The size values for the gray bubbles range from 4 to 10 => BRADIUSMIN=4pt BRADIUSMAX=10pt. Now, a size value x within the according range should result in a bubble of radius x pt. However, the red bubbles with sizes 8 and 6 are bigger than the corresponding gray bubbles with sizes 8 and 6. What am I missing here? Any help is appreciated! DATA test;     x++5;     y=5;     INPUT size_red size_gray; DATALINES; 20 10 8 8 6 6 4 4 RUN; PROC SGPLOT DATA=test;     BUBBLE X=x Y=y SIZE=size_red / BRADIUSMAX=20pt BRADIUSMIN=4pt FILLATTRS=(COLOR=RED) NOOUTLINE;     BUBBLE X=x Y=y SIZE=size_gray / BRADIUSMAX=10pt BRADIUSMIN=4pt FILLATTRS=(COLOR=lightGRAY) NOOUTLINE; QUIT; ... View more

Dear all, I have a macro (%foreign_macro) that needs to process a pretty big sas data set containing time series data ("very_large"). Unfortunately, with each execution the macro can only consider one specific day in the input data set (here: &input_data.=very_large). The macro is pretty fast for dates at the beginning of the data set, but the longer it has to search through the data, the longer it takes to execute. After the 80th iteration (and >300 to go), I had to stop SAS because it already took the macro several minutes to execute. In the example below, the macro %my_macro executes the aforementioned macro %foreign_macro for each of the dates given in the macro variable &dates. Uncommenting the code reveals the idea that I had for lowering the execution time. First, a duplicate of "very_large" called "_very_large" is made for use in further steps (disk space is not a restriction... yet). Second, before executing %foreign_macro, data that is no longer needed gets removed from the (duplicate) data set. This way, the macro doesn't have to go over this data time and time again with each execution. Unfortunately, when using modify & remove, it seems that data isn't actually (physically) removed. The execution time still increased with each iteration, and while the program was running "_very_large" never decreased in file size (it remained the exact same size as "very_large"). My first question would be if there's another way to remove the first observations of a data set (other than using modify & remove)? The second question is if there's a better solution for my problem (other than to rewrite %foreign_macro)? Thank you very much for your time. Any help is appreciated! Data very_large;     Attrib Date Length=4 Informat=date11. Format=date11.;     Attrib Vars Length=3;     Input Date Vars; Datalines; 09-JAN-2013 10 09-JAN-2013 20 09-JAN-2013 10 10-JAN-2013 20 11-JAN-2013 10 11-JAN-2013 20 Run; %Macro foreign_macro(     input_data=,     output_data=,     date= );     Data &output_data.;         Set &input_data.;         If Date>&date. Then Stop;         If Date=&date. Then Do;             Vars=Vars/2;             Output;         End;     Run; %MEND; %Macro my_macro(input_data=,dates=);     %Let dates_num=%eval(%sysfunc(count(%str(&dates.),%str(,)))+1); /*     * FIRST: work with a duplicate of very_large ;     Data _&input_data.; Set &input_data.; Run;     %Let input_data=_&input_data.; */      %DO i=1 %TO &dates_num.; /*          * SECOND: remove data from the duplicate that is no longer needed ;         Data &input_data.;             Modify &input_data.;             If Date>=%Scan(&dates,&i.,%Quote(,)) Then Stop;             Remove;         Run; */         %foreign_macro(             input_data=&input_data.,             output_data=output&i.,             date=%Scan(&dates.,&i.,%Quote(,))         );     %End; %Mend; %my_macro(     input_data=very_large,     dates=%str("09-JAN-2013"d,"11-JAN-2013"d) ); Nachricht wurde geändert durch: Jörg Honnacker ... View more