@BayzidurRahman I suggest you check the LIBNAME Statement: JSON Engine documentation and the ALLDATA= "name" LIBNAME Option   ALLDATA= "name" renames the ALLDATA data set to the specified name. A data set named ALLDATA is created by default when the JSON engine runs. The ALLDATA data set contains all of the information in the JSON file. This option creates the data set with the specified name instead. The new name must meet the conventions for SAS names. ... View more

I recommend you look at the documentation for PROC FORMAT  Here's a simple example using the PROC FORMAT CNTLIN data to create a custom format /* Create a PROC FORMAT cntlin dataset and a sample "have" dataset */ data work.cntlin work.have (keep=start) ; retain fmtName "$myCustomFormat" label 0 ; infile cards ; input start $10. ; output work.cntlin ; label+1 ; output work.cntlin work.have ; cards ; No Claims A00 0A0 5+ 3 4 00A 200 011 ; run ; /* Create character myCustomFormat using cntlin data */ proc format cntlin=work.cntlin ; run ; /* test myCustomFormat against sample "have" data */ data want ; set work.have ; want=putc(start,"$myCustomFormat.") ; put start= want= ; run ; ... View more

As @PeterClemmensen recommends please provide samples of both input datasets (table1 and table2), a sample of what you want in your output dataset based on the inputs, and the SAS code you have currently. For example: /* Create sample Table 1 data */ data work.table1 ; format date1 date7. time1 time5. ; infile cards ; input subject1 $3. date1 : date7. time1 : time5. ; output work.table1 ; cards ; 101 30Jan20 11:15 201 31Jan20 10:15 ; /* Create sample Table 2 data */ data work.table2 ; format date2 date7. time2 time5. ; infile cards ; input subject2 $3. date2 : date7. time2 : time5. ; output work.table2 ; cards ; 101 30Jan20 11:15 101 30Jan20 11:15 201 31Jan20 10:15 201 01Feb20 10:15 ; /* Create sample want data */ /* This is the resulting dataset I want */ /* Based on the input datasets table1 and table 2 */ data work.want ; format date date7. time time5. ; infile cards ; input subject $3. date : date7. time : time5. ; output work.want ; cards ; 101 30Jan20 11:15 201 31Jan20 10:15 201 01Feb20 10:15 ; /* My code to generate the want dataset from the input datasets table1 and table2 */ proc sql ; create table work.joined as select table1.*, table2.* from table1, table2 where table1.subject1=table2.subject2 and table1.date1=table2.date2 and table1.time1=table2.time2 ; quit ; ... View more

Hi @Dawa93  It would be helpful if you provided a "have" and "want" dataset (see other posts for examples)  Below I create a sample dataset (project) containing all possible combinations of v2, v5, v7 & v8 based on your IF logic. Then I included your sample code that creates the project_1 dataset. Finally, I've added my suggested code that creates project_2 dataset. Now given you have observations for each of the possible values of the input variables (assuming they can only have values of 0,1 and missing (.) you can easily review the output datasets (project_1 and project_2) against the logic and figure out what you need to adjust to get the results you are looking for.    /* Create sample data covering all possible values */ data project ; do v2=.,0,1 ; do v5=.,0,1 ; do v7=.,0,1 ; do v8=.,0,1 ; output ; end ; end ; end; end ; run ; /* Your example code */ data project_1; set project; *pandemic year*; if v2 = 1 and v5 = 1 then PandemicYear_1 = 1; *pandemic year 1; if v2 = 0 and v5 = 0 then PandemicYear_1 = 0; *none; if v2 = . or v5 = . then PandemicYear_1 = .; label PandemicYear_1 = Pandemic Year1; if v7 = 1 and v8 = 1 then PandemicYear_2 = 1; *pandemic year 2; if v7 = 0 and v8 = 0 then PandemicYear_2 = 0; *none; if v7 = . or v8 = . then PandemicYear_1 = .; label PandemicYear_2 = Pandemic Year2; run; /* Suggested code */ data project_2 ; /* You only need a single label statement */ label PandemicYear_1 = Pandemic Year1 PandemicYear_2 = Pandemic Year2; set project ; if v2 = 1 and v5 = 1 then PandemicYear_1 = 1; *pandemic year 1; else if v2 = 0 and v5 = 0 then PandemicYear_1 = 0; *none; else if v2 = . or v5 = . then PandemicYear_1 = .; /* need to catch potential values of v2 & v5 that are not covered above e.g. v2=1 and v5=0 */ else PandemicYear_1 = -999 ; if v7 = 1 and v8 = 1 then PandemicYear_2 = 1; *pandemic year 2; else if v7 = 0 and v8 = 0 then PandemicYear_2 = 0; *none; else if v7 = . or v8 = . then PandemicYear_1 = .; /* Should this be PandemicYear_2 */ /* need to catch potential values of v7 & v8 that are not covered above e.g. v7=1 and v8=0 */ else PandemicYear_2 = -999 ; run; ... View more

If you really need a single patient_id with multiple variables for each ICD value then you can do that in a single data step:  /* Set this to the maximum number of icdVar values in your icd_codes variable */ /* Note you may need to set this dynamically depending on your data */ %let icdVarCnt=13 ; data patient_icd; /* Keep the variables you want in the output dataset */ keep patient_id icd_var1-icd_var&icdVarCnt ; /* Create an array to hold the individual icd_var values */ array icd_var{&icdVarCnt} $100 ; /* Read test data */ input @1 Patient_ID $7. @8 ICD_CODES $600.; /* Loop through the icd_vsr variable extracting individual icd values */ do i=1 to &icdVarCnt ; /* extract individual icd values */ ICD_VAR(i) = scan(ICD_CODES, i, '|') ; /* Drop out off the loop if there are no more individual icd values */ if ICD_VAR(i) = '' then leave; end; /* write observation to output dataset */ output work.patient_icd ; datalines; 234423 (#1) Sickle-cell disease without crisis (D57.1) [POA=] | (#2) Idiopathic aseptic necrosis of unspecified bone (M87.00) [POA=] | (#3) Anxiety disorder, unspecified (F41.9) [POA=] | (#4) Osteonecrosis, unspecified (M87.9) [POA=] | (#5) Other symptoms and signs involving emotional state (R45.89) [POA=] | (#6) Other specified health status (Z78.9) [POA=] | (#7) Other specified counseling (Z71.89) [POA=] | (#8) Other long term (current) drug therapy (Z79.899) [POA=] | (#9) Encounter for therapeutic drug level monitoring (Z51.81) [POA=] | (#10) Counseling, unspecified (Z71.9) [POA=] ; run; Like others who have replied, I would not recommend this approach but only you know what your intended outcome is.  ... View more

Take a look at PROC MEANS and PROC SUMMARY proc means data=have n ; class id_appel demand_label ; var id_appel ; run ; ... View more

Can you provide more details on what you are looking to do? I'm assuming you want to subset some SAS dataset to only provide data where some variable matches todays date  I recommend you review the documentation on SAS Date, Time, and Datetime Values  Maybe this will help you on your way /* Create sample data set */ data work.have ; format date date9. ; do date="29MAR23"d to "05APR23"d ; if date=today() then value="I want these today" ; else value="I don't want these" ; output work.have ; end ; run ; /* Select observations where variable date is today */ data work.want ; set have ; if date=today() then output work.want ; run ; ... View more

See Data/ and Time functions    data _null_ ; datetime=datetime() ; datepart=datepart(datetime) ; day=day(datepart) ; month=month(datepart) ; year=year(datepart) ; hour=hour(datetime) ; put _all_ ; run ; ... View more

I recommend you review the documentation and search for Numeric Precision   ... View more

Run this example code and review the SAS log options mprint mlogic symbolgen ; %macro typ(cust=) ; %if &cust=abc or &cust=def %then %do ; %PUT *** TRUE - Cust is &cust *** ; %end ; %else %do ; %PUT *** FALSE - Cust is &cust *** ; %end ; %mend typ ; %typ(cust=abc) ; %typ(cust=def) ; %typ(cust=DEF) ; %typ(cust=ghi) ; SAS log, in spoiler  815 options mprint mlogic symbolgen ; 816 817 %macro typ(cust=) ; 818 %if &cust=abc or &cust=def %then 819 %do ; 820 %PUT *** TRUE - Cust is &cust *** ; 821 %end ; 822 %else %do ; 823 %PUT *** FALSE - Cust is &cust *** ; 824 %end ; 825 %mend typ ; 826 827 %typ(cust=abc) ; MLOGIC(TYP): Beginning execution. MLOGIC(TYP): Parameter CUST has value abc SYMBOLGEN: Macro variable CUST resolves to abc SYMBOLGEN: Macro variable CUST resolves to abc MLOGIC(TYP): %IF condition &cust=abc or &cust=def is TRUE MLOGIC(TYP): %PUT *** TRUE - Cust is &cust *** SYMBOLGEN: Macro variable CUST resolves to abc *** TRUE - Cust is abc *** MLOGIC(TYP): Ending execution. 828 %typ(cust=def) ; MLOGIC(TYP): Beginning execution. MLOGIC(TYP): Parameter CUST has value def SYMBOLGEN: Macro variable CUST resolves to def SYMBOLGEN: Macro variable CUST resolves to def MLOGIC(TYP): %IF condition &cust=abc or &cust=def is TRUE MLOGIC(TYP): %PUT *** TRUE - Cust is &cust *** SYMBOLGEN: Macro variable CUST resolves to def *** TRUE - Cust is def *** MLOGIC(TYP): Ending execution. 829 %typ(cust=DEF) ; MLOGIC(TYP): Beginning execution. MLOGIC(TYP): Parameter CUST has value DEF SYMBOLGEN: Macro variable CUST resolves to DEF SYMBOLGEN: Macro variable CUST resolves to DEF MLOGIC(TYP): %IF condition &cust=abc or &cust=def is FALSE MLOGIC(TYP): %PUT *** FALSE - Cust is &cust *** SYMBOLGEN: Macro variable CUST resolves to DEF *** FALSE - Cust is DEF *** MLOGIC(TYP): Ending execution. 830 %typ(cust=ghi) ; MLOGIC(TYP): Beginning execution. MLOGIC(TYP): Parameter CUST has value ghi SYMBOLGEN: Macro variable CUST resolves to ghi SYMBOLGEN: Macro variable CUST resolves to ghi MLOGIC(TYP): %IF condition &cust=abc or &cust=def is FALSE MLOGIC(TYP): %PUT *** FALSE - Cust is &cust *** SYMBOLGEN: Macro variable CUST resolves to ghi *** FALSE - Cust is ghi *** MLOGIC(TYP): Ending execution. ... View more

SAS is probably a lot larger than you realize  Here's a good starting place Programming Documentation for SAS® 9.4 and SAS® Viya® 3.5  ... View more

Although the code @Kurt_Bremser has supplied works for the example data provided. You also need to be aware that it ONLY works assuming your input data earliest observation has date="05APR2021"d  If you were to run this code against another dataset that which has a different value earliest value for date, e.g. 01APR2021 then week 1 is going to be 01APR2021- 07APR2021. If you want have consistent behavior where week 1 is always 05APR2021-11APR2021 then you are going to want to remove the PROC SQL step and hard code your starting date (&begin)  INTCK Function data have ; format date date7. ; do i=3 to 100 by 3 ; date="04APR2021"d+i ; output have ; end ; run ; proc sort data=have ; by date ; run ; data want; set have; week=intck("week","05APR2021"d,date,'c') + 1; run; ... View more

@Nietzsche I'm not sure I understand the problem, and like others, I'm not downloading or opening attachments. I find that the best approach is to simplify the question, and I suspect you may want to look at PROC SUMMARY (see example below) /* Create sample "have" data */ data have ; do sector ="A","B","C" ; do subSector = "1","2" ; do id=1 to int(ranuni(1)*3)+1 ; value1 = int(ranuni(2)*10) ; value2 = int(ranuni(3)*10) ; sumOfValues=sum(value1,value2) ; output ; end ; end ; end ; run ; proc summary data=have ; by sector subSector ; var sumOfValues ; output out=want sum=sumOfValue ; ; run ; If this isn't what you are looking for then I recommend you Simplify the question Provide sample input data in the form recommended by @PaigeMiller (See these instructions) Provide sample output data in the form recommended ... View more