Re: [boost] SQL: next iteration of sqlpp11

Subject: Re: [boost] SQL: next iteration of sqlpp11
From: Edward Diener (eldiener_at_[hidden])
Date: 2014-02-01 12:14:19

On 2/1/2014 9:04 AM, Roland Bock wrote:
> Dear boosters,
>
> after the lively discussion about sqlpp11 in November 2013
> (http://lists.boost.org/Archives/boost/2013/11/208388.php), sqlpp11 has
> evolved quite a bit and I want to give you an update.
>
> https://github.com/rbock/sqlpp11
>
> *) sqlpp11 now supports prepared statements.
>     For instance
>
>     auto ps = db.prepare(select(all_of(tab))
>                         .from(tab)
>                         .where(tab.alpha != parameter(tab.alpha)));
>     ps.params.alpha = 42;
>     for (const auto& row: db.run(ps))
>     {
>        std::cerr << "alpha: " << row.alpha << std::endl;
>        std::cerr << "beta: " << row.beta << std::endl;
>        std::cerr << "gamma: " << row.gamma << std::endl;
>     }
>
> *) Binding of parameters and results for binary transport is now
>     supported.
>     Both, parameters of prepared statements and results can be
>     bound to in-memory data of the database now.
>
>     In the example above, using the provided mysql connector, neither
>     parameters nor the non-text results are converted to/from string.
>
> *) Support for multi-row insert:
>     //First, you create an insert:
>     auto i = insert_into(tab).columns(tab.alpha, tab.beta);
>
>     //Then you add values (one or more times):
>     i.values(tab.alpha = 7, tab.beta = "seven");
>
>     // Then you run the insert
>     db.run(i);
>
>     There is no support for array binding in sqlpp11, since this is
>     highly vendor specific afaict. But it would be fairly easy to
>     add support for this in the vendor-specific connector library. I'd
>     be happy to support you if you wanted to give it a try.
>
> *) New interpreter for SQL expression trees:
>     This has changed and improved a lot. sqlpp11 has been stripped
>     of all vendor specific code and will serialize standard SQL.
>     The connector libraries can specialize an interpreter template to
>     change the representation of certain aspects. For instance, mysql
>     would represent the concatenation of strings as CONCAT(a,b)
>     instead of a||b.
>
>     But of course, the connector libraries are not required to turn
>     the expression into a string. They can interpret it any way they
>     like. I strongly believe that this is going to allow for writing
>     connectors to XML, input streams, maps, etc. (think LINQ).
>
> *) I added a sample code generator (in python) that generates table
>     representations from a DDL
>
> Todo-List:
> ----------
> *) More thorough documentation
> *) More thorough and easier to comprehend tests
> *) Some of you mentioned that Boost.Fusion or Boost.Preprocessor
>     might do be able to do the job of code generation very elegantly.
>     I won't do that myself
>     but I'd be interested to see such solutions.
>     Since code generation is not integral part of the library
>     it would make sense to have such solutions as support libraries.
> *) Look into cursors
>
>
> FYI, I am going to give a talk about sqlpp11 and some template
> patterns used inside of it on February 27 in Munich, see
> http://www.meetup.com/MUCplusplus/events/160688442/
>
> Please let me know your questions/thoughts/suggestions/rants.
> Contributions welcome, of course :-)

Does not build under Windows.

You need documentation somewhere.