You, good SAS user, are a genius!  I typically use a CALL EXECUTE in a DATA _NULL_ to control iterative variables but in this case I moved it inside.  Thank you for your helpfulness! ... View more

To further clarify, I have found a workaround by copying/pasting the code generated in the LOG, so I'm not trying to find alternative solutions. I'm just dying to know why PROC SQL is having an issue processing the %DO loop generated statements.  We use this technique frequently in our code and have not had this issue before, so we're just curious if anyone knows why.  Thanks! ... View more

I appreciate the input--this is a multi-hundred gigabyte permanent dataset and each operation on it takes roughly an hour or more--a single PROC SQL with windowed variables has proven, over our work with our nearly 20 TB of data, to be the most efficient approach. We have models that we build which will have starting variable cardinality in the thousands, so the number of variables is not the issue, though the feedback is appreciated. If anyone has an answer as to why the QUERY in the log that is generated as a result of the macro works perfectly as is, but the macro fails, that would be most appreciated.  Thank you. ... View more

These are windowed variables to be passed into a feature selection algorithm for creating a logistic model. Do you want more of a partial log than what is at the bottom of the post, or of a different part of the log?  If so, here is some of it: MPRINT(ONE_YEAR_RU):   proc sql; NOTE: Line generated by the macro variable "VALUE". 1        num_interactions_30                           --                           22                            --                            76 MPRINT(ONE_YEAR_RU):   create table pb_sales_ru as select party_id, sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_interactions else 0 end) as num_interactions_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_trans else 0 end) as num_trans_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_rtrns else 0 end) as num_rtns_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_store_trans else 0 end) as num_store_trans_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_store_rtrns else 0 end) as num_store_rtrns_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_dir_trans else 0 end) as num_dir_trans_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_dir_rtrns else 0 end) as num_dir_rtrns_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then items_purchased else 0 end) as items_purchased_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_store_items else 0 end) as num_store_items_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then num_dir_items else 0 end) as num_dir_items_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then net_sales else 0 end) as net_sales_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then net_store_sales else 0 end) as net_store_sales_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then net_dir_sales else 0 end) as net_dir_sales_30 , max(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then if1 else 0 end) as if1_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then iq1 else 0 end) as iq1_30 , max(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then if2 else 0 end) as if2_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then iq2 else 0 end) as iq2_30 , max(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then if3 else 0 end) as if3_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then iq3 else 0 end) as iq3_30 , max(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then if4 else 0 end) as if4_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then iq4 else 0 end) as iq4_30 , max(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then if5 else 0 end) as if5_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then iq5 else 0 end) as iq5_30 , max(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then if6 else 0 end) as if6_30 , sum(case when tran_date between '13OCT2013'd-30 and '13OCT2013'd then iq6 else 0 .... {TONS MORE LOG...} .... ERROR 22-322: Syntax error, expecting one of the following: ',', LEN, LENGTH. ERROR 76-322: Syntax error, statement will be ignored. ... View more