From bmaclean@vcommerce.com Mon, 30 Jul 2001 13:05:11 -0700 From: Brian MacLean Date: Mon, 30 Jul 2001 13:05:11 -0700 Subject: RE: 2 Gb file limit problem Message-ID: MIME-Version: 1.0 Content-Type: text/plain How about using a pipe and the split command if necessary?   If your unix machine can handle files greater than 2 gig's but just SQL*Plus can't, then:   1) mknod pipe.lst p 2) cat pipe.lst >mybigfile & 3) sqlplus u/p         spool pipe.lst         @my_commands.lst                spool off         exit 4) rm pipe.lst   If the unix machine also has a 2 gig limit, then:   1) mknod pipe.lst p 2) split -b 1073741824 pipe.lst mybigfile_ & 3) sqlplus u/p         spool pipe.lst         @my_commands.lst                spool off         exit 4) rm pipe.lst   These examples should give you enough to work with.        -----Original Message-----From: JOE TESTA [mailto:JTESTA@longaberger.com]Sent: Monday, July 30, 2001 12:52 PMTo: Multiple recipients of list ORACLE-LSubject: Re: 2 Gb file limit problem how about this:   (avg_row_size + delimiters) * number_of_rows = total bytes.   total bytes / 1900000000 = number of pieces.   number_of_rows / number_of_pieces = number of rows per piece   select number of rows needed multiple times, spooling each one individually.   then sqlldr all the pieces.   joe   >>> Satish.Iyer@ci.seattle.wa.us 07/30/01 02:20PM >>> Hi List,   I need to transport few tables from one instance to another and of course found the sqlldr method much faster than the exp/imp. But the problems is for large tables .When I spool such input tables to a flat file , it stops spooling into it after about  2 Gb. Any possible solutions to get around it. I am on AIX 4.3.3/8.1.5   My ulimits on AIX  are time(seconds)        unlimitedfile(blocks)         unlimiteddata(kbytes)         unlimitedstack(kbytes)        unlimitedmemory(kbytes)       unlimitedcoredump(blocks)     unlimitednofiles(descriptors) 2000   Thanks   Satish