Take a look at this very good resource with nicely explained examples:   http://www2.sas.com/proceedings/sugi29/265-29.pdf ... View more

This is not proc report, this is part of a data step that creates text output. In this case, it is used to create MIME/HTML/XML code "manually", without using SAS ODS or other helpers. Since important parts of the code are missing (the start of the macro definition, for instance, and parts of the data step), it is not possible to thoroughly help you. One thing that stands out is a run; statement right in between, followed immediately by another put; sequence. This can't work, as you need a new DATA statement after the step boundary caused by the run; To further understand the code, one would also need the dataset that is used as base for the data step.     ... View more

ODS Tagsets is used for exporting data, not importing data. Proc Import and Libname methods may work, assuming you have the correct license, in this case SAS/Access. ... View more

To get MEANS/SUMMARY to generate all cells you need to include the COMPLETETYPES option. proc summary data=demog(keep=race trt) nway  completetypes;   class race trt/preloadfmt ;   format race race. trt trt.;   output out=summary ; run; ... View more

Hi: ACROSS items can't be computed. So they have to exist before the PROC REPORT starts. The fact that TREATMENT wasn't in your data was one of the reasons I was of the thinking that your posted code might not have produced the output you posted.   Since you did NOT show your ODS RTF statement and you want to show all the border lines in your report, I would recommend using the default RTF style -- something like: ODS RTF file='test_report.rtf' style=RTF;   Did you experiment with the spanning header technique I showed without creating a TREATMENT variable to see how that might work for you. You do NOT need a TREATMENT variable to make a spanning header. column category _name_ ('Some Long String' (a b c d overall));   To do what you want with Min_Max then you need to use a COMPUTE block with an IF statement or you need to pre-process the data for PROC REPORT and change the value before you send the data to PROC REPORT.   cynthia ... View more

Hi: There are a couple of different ways you could do this. One way is to create a custom style template. The other way is to use statement level style specifications in the TITLE statement or the PROC REPORT statements. It really depends on whether you need to do this for multiple reports or whether you need it for an ad-hoc type report.   There are several papers about style templates such as this one http://support.sas.com/resources/papers/proceedings10/033-2010.pdf and several papers about using PROC REPORT and statement level style overrides like these: http://support.sas.com/resources/papers/proceedings13/366-2013.pdf http://www2.sas.com/proceedings/sugi30/244-30.pdf   I'd recommend starting with the style override method first and then work on style templates once you get the hang of style elements and attributes and what has an impact on what piece of the report. cynthia ... View more

As mentioned you can generate LOTS of output; But PUT is conditional so if you are looking for specific values or combinations of value or missing values you can selectively report on them. A brief example: data new;      Infile datasource;      input <input a bunch of variables from the data source>;      if DateOfBirth > DateOfService then put "ERROR: Date of Birth after Date of Service for ID " Idvar dateofbirth dateofservice; run; So I get a message in the log that we have some error related to two dates and identifying information. Added benefit, if you start the message with ERROR: or WARNING: in upper case with the colon then the text in the log will appear with the colors you have set for error or warnings, default red and yellow respectively. ... View more

No need Macro. data _null_; set sashelp.vmember(where=(libname='WORK' and memtype='DATA')) end=last; if _n_ eq 1 then call execute('proc datasets library='||libname||' nolist nodetails; change '); call execute(strip(memname)||'=want'||strip(_n_)); if last then call execute(';quit;'); run; Xia Keshan ... View more

Alternatively, you can use the line statement in compute block in proc report syntax: compute before _page_; line "this is the title"; line ''; endcomp; Thanks, Jag ... View more

proc sort data=have; by city; run; data want(keep=city unique_count); set have; by city; lag_id=id-lag(id); if first.city then unique_count=1; if not first.city and lag_id ne 0 then unique_count+1; if last.city then output; run; ... View more

data want;      merge vitals (in=a) dosing (in=b);      by PATIENT;      if a and b; where amt ne ' '; run; thanks ... View more

chisq-  is for the independent test of row and column (each cell at least 5 frequency  ). fisher exact - is for the small sample data ( one of cell has less than 5 frequency ). CMH- is for multiple contingency table analysis . Why not check SAS documentation firstly ? ... View more

Post an example of the code you tested and a specific question about it. A simple iterative DO loop like in your original question work the same as in any other language. The DO statement in SAS actual has many variations and can be extremely powerful, but without more details of what you are trying to do it is hard to give specific advice. Here is link to on-line document for the iterative form of DO. SAS(R) 9.4 Statements: Reference, Second Edition ... View more