patstat | amazing scripts , mainly in SQL , made by members | Database library

&:$29 is probably your problem. & there is telling SAS to wait for two consecutive delimiters (this is one way of dealing with delimiters embedded in the data, though it is not a common one). That's not (apparently) correct for your data, as far as I can tell.

Remove the & and see if it imports correctly.

Fix the import to correctly ready in apln_nr_original.

Note the bolded sections below.

So I counted out the number of variables and I think it's the 8th variable, which looks like it should be, AT2007/000486 from the record. However, SAS has it as ., which means it thinks it should be a numeric, when it's actually a character variable. So you need to modify your code to account for that. I'd suggest how to do that but you didn't include any code.

NOTE: Invalid data for appln_nr_original in line 5286 53-65. RULE: ----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+--- 5286 6697,AT,2007000486,W ,2007-10-17,2007,WO2007AT00486,AT2007/000486,PI,0,Y,N,N,2006-12-22, 89 2006,1110640,2008-07-03,2008,6698,0,38109624,4532,10,2,2,1 146

appln_id=6697 appln_auth=AT appln_nr=2007000486 appln_kind=W appln_filing_date=2007-10-17 appln_filing_year=2007 appln_nr_epodoc=WO2007AT00486 appln_nr_original=. ipr_type=PI

You need to start SAS with unicode support to read UTF-8 characters.

You could try setting ENCODING=ANY on the INFILE or FILENAME statement in your current SAS session. The encoding shouldn't matter for the numbers. But if you really have UTF-8 characters that cannot be transcoded into single byte WLATIN1 characters then you will probably have trouble working with those strings.

I found a way to add the citation table to my query.