John Gehring wrote:
> I am adding records to my MySQL database that contains a column of type
> 'mediumtext' (call it ColA). The amount of data is roughly 75K bytes. If I
> use the SELECT ... INTO OUTFILE ... statement to get the data, I
> successfully get all the data of the record. However, if I use a select
> statement within an embedded application (using mysql++), the data from ColA
> is truncated at 65K bytes.
Things to check:
- What character set are you using? Unicode character representations
may do surprising things when processed by naive code. See chapter 6 in
the user manual.
- "65K" bothers me. Are you using SI suffixes here -- 65,000 bytes --
or do you mean that it's actually getting over the special 64K boundary?
- Are you sure you don't need a BLOB column here? Or, at least treating
it with the precautions due a BLOB column? See examples/load_jpeg.cpp
and examples/cgi_jpeg.cpp.
> I see quite a bit of documentation for the mysql
> embedded library related to binding data for result sets. But I don't see
> anyplace in the mysql++ code that uses that functionality (like
> mysql_stmt_bind_result()).
Yes, but there's no requirement to use those APIs. MySQL++ takes the
traditional approach: it uses text representations of data, in plain old
SQL forms. You must use BLOBs to get bit-for-bit accuracy, and even
then, it goes across the wire in escaped text form.
> Am I missing something, or does the mysql++
> library have a limit on how large a field of data can be returned from a
> select statement?
MySQL++ imposes no limits above what the underlying C API does. That's
not to say that you can't abuse the library to make it truncate data,
just that there are proper ways to get the data without truncation.
