Hi all, is there a way in SAS to find the second last value of a column quicker than with something like proc sql;   create table &_output. as   select        a.id,        a.version,        a.value_to_be_looked_up   from        abc.zzz a   join   (      select id,  max(version)-1 as version_before      from      abc.zzz       group by      id   ) b   on   (   a.id = b.id   and   a.version = b.version_before )   where   something = some_value   ; Thanks a lot, th ... View more

Hi all, strange issue here and I have no idea what's going on. Any hint would be highly appreciated. A SAS job that runs fine during the night (hundreds of times already) runs once a month also during the day. This one time it loads 120.000 rows into the DB/2 instance, while during the night there are about 60.000 rows to process. So when it comes to the daily run the job connects to DB/2 and...does nothing. It does neither loa nor update nor does the job itself crash with an error. It just idles on the DB without throwing any error on SAS or DB/2 side. Has anyone experienced such an issue before? Please note, that this job connects to the very same DB and table in the night and runs fine. If the row number would be problem I would assume that other jobs, that do similar stuff with similar row number also crashed, but they run fine. Thanks a lot for any idea, Thomas ... View more

Hi all, I work with SAS DI Studio 4.2. and I am currently reading through some high level papers on SAS Performance. These are talking about increasing SQL Join performance by ordering columns the right way - so, what IS the right way to order columns? Let's say there are 2 tables each containing 80 columns and I need to join them by 3 columns #2, #44 and #78. Do I change #2, #44 and #78 to columns #1,#2,#3 before joining? Another question is about "in job parallelization" - let's say there is 1 Job containing 1 source table, 2 Extracts based on the source table, 1 Append that appends the results of the 2 Extracts and a Table Loader that writes the data into another table. How would I - make the 2 Extracts work in parallel via DI Studio - make this work without blocking the source table Thanks a lot, th ... View more