WHERE is used only to select observations from the input and skip the not required. IF can be used to select observations for output but in most cases to select observations for more specific programing. In some cases it may be easier to code and use: SELECT statement with WHEN.  Check documentation for examples. Using SQL you have CASE clause - similar to the sas datastep SELECT.      ... View more

If I understand you correctly then you need to add the appropriate condition: if subject = 1010 and alive = 'No' and missing(score) then score = 100; ... View more

@Phil_NZ - a little bit of theory - what is a file: You probably know, at least the next two terms: VARIABLE and RECORD. SAS uses the terms OBSERVATION (instead record). LRECL stands for Logical RECord Length.  RECFM stands for RECord ForMat.   If you sum the length of all varaibles in a record, including delimiters and control characters (like EOL = End Of Line) you get the Record Length. Depending on record format - the record length may be a fix size or variable size. When reading a text file sas writes to the log the minimum size and the maximum size of the input record.   Data is written to Hard-Disk as a block of records. The total length of a block is called BLOK SIZE.   As that, LRECL=1/32767  is erroneous. LRECL can get an integer value which is dependent on the file you read/relate to. You need know the real length of the record when the file was written.   In some cases each block contains one record only - then LRECL is equal the Block Size. I think it happens when you write, for example, to a printer.   In your case, using PROC IMPORT or a data step with INFILE, you need not define the LRECL.   ... View more

Post first few lines from your csv file. Does the first line contain the variable names? Are those names sas valid names?   What happens if you assign option firstobs=2 ? ... View more

You input test tables are each 9 observations. By comparing each possible pair of the two tables observations the output contains more observations then you posted as wanted: data table1; infile cards dlm='0920'x truncover; input ID $ A1 B1 c1 $; cards; a 0.12 30 promo1 b 0.12 30 promo1 b 0.15 30 promo1 b 0.12 30 promo1 b 0.18 40 promo2 c 0.12 30 promo1 c 0.18 40 promo2 c 0.165 40 promo2 c 0.18 40 promo2 ;run; data table2; infile cards dlm='0920'x truncover; input ID $ A1 B1 c1 $; cards; a 0.12 30 promo1 b 0.12 30 promo1 b 0.15 30 promo1 b 0.12 30 promo1 b 0.18 40 promo2 c 0.12 30 promo1 c 0.18 40 promo2 c 0.165 40 promo2 c 0.18 40 promo2 ;run; proc sql; create table want as select a.*, case when (a.a1 ne b.a1) then 1 else 0 end as flag from table1 as a left join table2 as b on (a.ID=b.ID and a.c1=b.c1 and a.b1=b.b1); quit; AS for the first compare method mentioned:         ... View more

@Phil_NZ relating to your's  "Because you wrote down that <variable(s)> so I expect that I can list all unique values of 2 or more variables at one time but when running in SAS, I get error. Other than that, these two solutions work well with one variable at one time." 1) As already said, any list used in sql need be separated by comma. 2) The meaning of select distinct <var1>,<var2> [,...,varn] is to get      the distinct combinations of the variables in the list. 3) To get distinct value of each variable in a separate output dataset you can use: proc sql; create table sex as select distinct sex from <dataset>; create table age as select distinct age from <dataset>; quit; ... View more

proc sql; create table want as select distinct <variable(s)> /* SEX or AGE */ from have; quit; or  proc sort data=have(keep=<variable> out=want NODUPKEY; by <variable>; run; ... View more