Auto replying for someone who eventually has the same problem as i did! The solution i found is to put the following code at the biginning of the SP code: [pre] %global _result; %let _result=STATUS; [/pre] This way only the status of the stored process is outputed and all other outputs are supressed. ... View more

When we ask for simplification what we need is essential code - the part that you think is giving you the problem. That helps us home in on the true issue. I suspect that there is more than one. A couple potential ones have already been identified. Here is a 'simple' format that may answer a couple of the questions. Notice the KEEP= option and especially the use of the HLO variable to allow a nested format. Finally I included some dummy data to test the format. [pre]data intervals(keep=fmtname start end label hlo); retain fmtname 'myfmt'; start = '12jan2000:01:01:01'dt; end = '24nov2005:11:12:13'dt; label = 'datetime18.'; hlo = 'F'; output; start=.; end=.; hlo='O'; label= '.' ; output; run; proc format cntlin=intervals; run; data tryit; value= '10jan2000:01:01:01'dt; put 'Below range ' value= datetime. value=myfmt.; value= '10jan2004:01:01:01'dt; put 'In range ' value= datetime. value=myfmt.; value= '10jan2006:01:01:01'dt; put 'After range ' value= datetime. value=myfmt.; run;[/pre] When writing a format using a control data set it is sometimes helpful to know some of the values that can be taken on by the supporting variables, such as HLO. I sometimes write a dummy format using the VALUE or INVALUE statement and then print out the data set: [pre] proc format ; value testdt low- < '01jan2011'd = 'early' '01jan2011'd - '01jan2012'd = [date9.] other = 'unk'; run; proc format cntlout=testdt(where=(fmtname='TESTDT')); run; proc print data=testdt; run; [/pre] Art ... View more

> alternatively examine the number of delimiters in > the _infile_ buffer > not good enough if the delimiter might be embedded in > data, like in "erty|tyu" in: > 2345|3456|"erty|tyu"|4567|| How about counting the words. [pre] 1951 filename FT15F001 temp; 1952 canWeCountWords: 1952! data _null_; 1953 infile FT15F001; 1954 input @; 1955 c = countW(_infile_,'|','MQ'); 1956 put 'NOTE: ' c=; 1957 list; 1958 parmcards; 1961 ;;;; NOTE: The infile FT15F001 is: (system-specific pathname), (system-specific file attributes) NOTE: c=6 RULE: ----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+----9----+----0 1 2345|3456|"erty|tyu"|4567|| 27 NOTE: c=2 2 | 1 NOTE: 2 records were read from the infile (system-specific pathname). The minimum record length was 1. The maximum record length was 27. NOTE: DATA statement used (Total process time): real time 0.01 seconds cpu time 0.01 seconds [/pre] ... View more

I really wasn't seeing the full picture, specially that part of the index being rebuilt at every insert. I now use indexes to delete or update records that match and afterwards, i drop all indexes, append the nonmatch records and rebuild the indexes. This solution as droped my processing time to less than 1/5th of the original time, which is very acceptable considering this was about the first load of the master table, and the next loads will have a much shorter range of data. Thanks for all the tips. Best regards, Pedro Azevedo. Message was edited by: dropez ... View more

If the machines running Enterprise Guide can access the location where the SAS code for the STP is stored, your best option may be to simply include the SAS code into the process flows when needed. You may have to make small modifications to the code to get it to behave properly in Enterprise Guide. If so, save a copy and include that into the process flows. Vince DelGobbo SAS R&D ... View more

depending on the size of small, you might achieve a hash join in SQL. Paul Kent's TS note about SQL iirc describes how the buffersize option increases the size of the small table that can be accomodated. http://support.sas.com/techsup/technote/ts553.html .. ... View more