Re: Error: "Out of memory while reading tuples." in pushing table from SAS to PostgreSQL on Mac

On 15/12/14 04:44, Andy Colson wrote:
> On 12/13/2014 10:03 PM, wetter wetterana wrote:
>> Hi,
>>
>> I'm passing rows from SAS to PostgreSQL (I assign a libname and use a
>> PROC APPEND).  This works fine with smaller tables (~below 1 million
>> rows).  However, as tables get larger I receive the following error
>> messages:
>>
>>
>> "ERROR: CLI describe error: Out of memory while reading tuples.; No
>> query has been executed with that handle"
>> and
>> "GLOBAL SYSDBMSG POSTGRES: Out of memory while reading tuples.; No
>> query has been executed with that handle
>> GLOBAL SYSDBRC HY000"
>>
>> I've tried to change memory settings on the PostgreSQL server, but
>> can't solve the problem.  As far as I could understand—I'm new to
>> PostgreSQL ;)—it seems that PostgreSQL want to somehow read
>> information on the whole table before processing it and this behavior
>> could eventually be switched off, but I might be wrong here.
>>
>> FYI:
>> - I run SAS 9.4 on a windows machine.
>> - I run PostgreSQL server on a MAC: PostgreSQL 9.3.5 on
>> x86_64-apple-darwin12.5.0, compiled by Apple LLVM version 5.1, 64-bit
>>    I've Pgadmin 1.18.1 installed.
>>
>> Any help would be much appreciated!!
>>
>> THANKS!
>>
>>
>> PS: For several reasons, I cannot use the bulkload feature in SAS for
>> this job.
>>
>
>
> That error is coming from SAS, not PG.  SAS must have pretty bad
> documentation because when I google "GLOBAL SYSDBRC HY000" there is
> pretty much nothing.  (Also, I've never used, or even heard of SAS.
> At first I thought you meant serial attached scsi)
>
> Does SAS support a cursor of some kind?
>
> -Andy
>
>
SAS: Statistical Analysis System (this is the original name I remember,
I think they have renamed it)

SAS: Special Air Service (elite British fighting force, deliberately
misnamed to confuse the Germans in World War II)

Funny, just now I thought of the second definition first, but obviously
the first applies.

I actually used SAS many years ago, long before I had written any SQL -
let alone directly used any database!


Cheers,
Gavin