Hi,

I have a table that contains a BLOB and I'd like to use an SSQLS to
query it.
e.g.
Int
Int
Blob - containing 200 floats in binary format
Varchar(128)
Varchar(128)

I use an sql_create_5() statement to define my table elements and then I
do a query.storein() on the vector created with the structure (which in
turn is created by sql_create_5()). Everything works fine. Both Int and
Varchar fields are correct!

The problem is that - if I understood things correctly - I can only
define a BLOB as a std::string. And when I execute my query and try to
access the data in the BLOB, it stops after the first hex 00 (which is
"normal" since it's a string). Problem is that even if I access the
data() member of the string, I only get the bytes till the first 00.

So basically my question is: Is it possible to use sql_create_#() with
BLOBS or do I have to go the "standard" way (non-SSQLS way that is)?

Hope I'm just missing something obvious and I'll be able to use SSQLS
since that feature really *rocks*.

Keep up the good work,

Viktor STARK



--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Mon, Nov 28, 2005 at 12:38:08AM +0100, Viktor STARK wrote:
Just to make sure, you've checked the size() of the string, right?

I have a bad feeling that somewhere the data is being assigned with
operator=(), but I don't know for sure. This would take some time,
delving into custom.pl.

- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Sun, Nov 27, 2005 at 08:51:30PM -0500, Chris Frey wrote:
Looks like a limitation in the conversion from Row.at() into ColData, which
takes only a char* in the constructor.

I don't know if Warren has fixed this in his private/devel tree or not.
This sounds familiar to me and I don't know why. :-)

Might take me a while before I can write a patch for this, given
time constraints. Should be easy to do, if someone wants to tackle it.
Look at Row::at(), the ColData template, and the const_string used in the
ColData typedef. Then add an operator const std::string&() to ColData.

- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

Viktor STARK wrote:
This truncation bug may be fixable.

But, don't expect it to be possible to define an SSQLS with "200 floats
in binary format" and have the one BLOB field be busted out into those
200 individual fields. SSQLS requires a 1:1 relationship between fields
in your structure and fields in the database.

I've put your truncation bug on the Wishlist. Since it's your itch,
you're in the best position to scratch it. I don't have any idea when I
can get around to it.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

Chris Frey wrote:
I don't recall anything about it. I presume that you're wanting to add
an std::string overload. There may be a reason it can't be done. We've
run into char* vs. std::string vs. SQLString vs.... problems before.

FYI, the only thing I have that's not checked in right now is my
investigation into Bakefile. It's marginally working, but not so much
that I'm happy with it yet. I will probably release it on the list as a
patch to the v2.1 branch first, for wider banging-on.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Mon, Nov 28, 2005 at 03:15:15PM -0700, Warren Young wrote:
This triggered a cool idea... it should be possible to write a class to
use in SSQLS, in place of std::string, that takes a std::string in the
constructor.

Of course, after I fix the ColData size bug. :-)

- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Mon, Nov 28, 2005 at 03:18:22PM -0700, Warren Young wrote:
Must be my imagination.


Should be fixable. Might not maintain ABI compatibility though.

The main problem is that data/size of raw field data is maintained right
up to the Row::at() conversion, where suddenly ColData has no notion of
anything but a char* in the constructor.

- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

Chris Frey wrote:
Yep, that's Worthy. It's on the Wishlist for post-v2.1.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Mon, 2005-11-28 at 15:15 -0700, Warren Young wrote:
into floats after the query.

Thanks.

Viktor STARK


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Mon, 2005-11-28 at 17:30 -0500, Chris Frey wrote:
this.

BTW, I've made some progress about the ColData size bug. I hope to come
up with something useful this WE.
As for writing the new class, mysqlpp:ate might be a good source of
inspiration, IMHO.

Viktor STARK


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw