From: Warren Young
Date: March 23 2011 9:08am
Subject: Re: StoreQueryResults destructor runs without end...
List-Archive: http://lists.mysql.com/plusplus/9265
Message-Id: <64C9DA60-CB21-4BD5-8D5F-716B2F49718D@etr-usa.com>
MIME-Version: 1.0 (Apple Message framework v1082)
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: quoted-printable

On Mar 23, 2011, at 3:00 AM, Warren Young wrote:

> On Mar 22, 2011, at 6:48 PM, Paul Dalach wrote:
>=20
>> I am pulling ~80k objects with four fields.  Does that constitute =
huge?
>=20
> Well, it depends on how big those fields are, doesn't it?  If they are =
four 32 byte strings, the amount of data in use is at least 10 GB.

Sorry, got my MB confused with my GB there.  80,000 records x 4 x =
32-byte strings is ~10 MB, not 10 GB.

Still, the answer may still be valid if the field sizes are large =
enough.

Keep in mind also the behavior of std::vector, which StoreQueryResult is =
based on.  (More or less, it is a 2D vector of mysqlpp::String.)  It =
doubles in size each time you exceed its size, so it can keep growing.  =
That means lots of RAM copies for such a large number of fields.

You can solve some of that by calling res.reserve(N), where N is larger =
than the number of records possible from the query, but you still have =
to deal with the fact that you're trying to store all that in RAM at =
once.=