Thanks it worked. ... View more

Hi, I have a question on IN operator. I'm using in operator but not  getting all the values. e.g   if State in ('NJ', 'New Jersey');     then I'm getting only NJ in the output and not New Jersey. Can you suggest a better way to get both of the values.   Thank you M ... View more

@Meng_Xu Please post your question as a new thread. This one is more than 5 years old. Please include your code and log. ... View more

Hi,   I found a pretty cool add-in to Base SAS on www.tatesoft.com that may be very usefull to you. The tool called Log Analyser is integrated into Base SAS and can produce Log summary report in a second. You may also analyse log in details if you wish. Really great tool! ... View more

Could you explain further? The example doesn't seem complete? The combination 20-22 exists in both IDs and there is no total. 20-20 is not a combination in either ID. PG ... View more

Something like this, untested code, PROC SQL; CREATE TABLE abc AS SELECT a.*, b.join_date, c.term_date FROM a, b, c WHERE a.id=b.id AND a.id=c.id; That will do an inner join keeping all the columns from a and just the respective dates from b and c. A left join would be similar; you can find the specifics in the SQL manual. ... View more

Suggest starting with the SAS-hosted documentation at SAS support http://support.sas.com website, along with additional technical / conference material on the site as well. Scott Barry SBBWorks, Inc. Suggested Google advanced search argument, this topic / post: cluster analysis site:sas.com ... View more

Sorry.I found some problem in my code; The following is corrected code. [pre] data i; input ID $ ADATE : MMDDYY10. SERVICE $; format adate mmddyy10.; datalines; 1 12/7/2009 I 2 1/12/2009 I 2 1/13/2009 OP 3 6/5/2009 I 4 7/29/2009 LA 5 3/23/2009 OP ; proc sort data=i nodupkey; by id adate service; run; data index; set i; where service not eq 'I'; run; proc sql feedback; *createing table contained observations which is needed to change into I; create table flag as select a.* from index as a left join i as b on a.id eq b.id where ( b.adate eq a.adate ) or ( b.adate eq (a.adate - 1)) group by a.id,a.service having count(*) ge 2; quit; proc sort data=flag nodupkey; by id adate service; run; data flag; set flag; _service='I'; drop service; run; proc sort data=i; by id adate; run; data result; merge i flag; by id adate; run; data result; set result; if _service eq 'I' then service=_service; drop _service; run; ... View more

Thank you for your suggestions. It was very helpful. Thanks, Sandhya. ... View more

Thank you. Sandhya. ... View more

since you want to avoid sorting, I presume your data are not in an order useful to the query. Consider a hash table. How wide and how many are the keys? If the keys of table1 can be stored with indexes on var1 and var2, then you should be able to collect your query results with a single pass through each table. However, sort/merge might take less time to get results if you have to addin the learning curve. good luck peterC ... View more

Thank you. The suggestions really helped. I used proc sql and used a sub-query. Thanks once again. Sandy. ... View more