One thing to try is to not use the DSD option on the FILE statement.
This will mean that you are not protected from embedded delimiters or unbalanced quotes. So you need to check your data before writing it.
It also will means that missing (aka NULL) values will be represented as a single character instead of by adjacent delimiters. So you need check to how Netezza will lines like this:
value1| |value3
instead of like this:
value1||value3
A lot of databases will allow you to include the work "null" (or perhaps some other character string) to indicate missing values. If so you might be able to use user defined formats to fix that.
data have;
length yr 8 make model $20 description $100 ;
yr=1997;
make='Ford';
model='E350';
do description = ' ',"Super, luxurious truck","Super, ""luxurious"" truck" ;
output;
end;
call missing(yr);
output;
run;
proc format ;
value null low-high=[best32.] other='null';
value $null ' '='null' other=[$200.];
run;
filename csv temp;
data _null_;
file csv dlm='|' ;
set have;
format _numeric_ null. _character_ $null. ;
put (_all_) (:) ;
run;
data _null_;
infile csv ;
input;
list;
run;
RULE: ----+----1----+----2----+----3----+----4----+----5----+----6----+----7
1 1997|Ford|E350|null 19
2 1997|Ford|E350|Super, luxurious truck 37
3 1997|Ford|E350|Super, "luxurious" truck 39
4 null|Ford|E350|Super, "luxurious" truck 39
