Re: Standard DBI Proposal
- From: Kevin Kenny <kennykb@xxxxxxx>
- Date: Wed, 31 Oct 2007 10:16:41 -0400
tunity5@xxxxxxxxx wrote:
As I said in my email, I did try to build [tclodbc] on non-Windows systems.
Again, it was possible a few years ago on Linux. My experience with
it says it is quite an accomplishment if you can get it to compile
successfully today, if at all.
OK, tclodbc appears to have toolchain problems on Unix. Noted.
Ahh! Now I see why you have mentioned ODBC performance a few times.
You are transferring large amounts of data back and forth between the
database server and the client. And it would seem that ODBC was not
performing because of all the data transfers back and forth.
I do not think so. You will never have enough memory for your data.
This is one thing I do not like about Sqlite interface. If you are
retrieving plain data rows from the server and doing all the work on
the client (joins, merges, etc.), please read up on databases and
SQL. Or buy our products :-) Otherwise, you are re-implementing what
a database server does.
When I've been forced into large client-side operations, it's been
because I've been dealing with multiple servers with intransigent
DBA's. When you have some of your data in Oracle, some in SQL Server,
and some in MySQL, and none of the DBA's will allocate you space
to load the other one's data, you're kind of stuck. Perhaps you're
telling me that people in that situation should go find a better grade
of customer? You haven't stated an organizational affiliation,
so it's hard to tell what products you're recommending. But a data
warehousing solution is a bit of overkill for a one-off conversion.
If it can let me connect to databases on non-Windows systems, I am
really looking forward to it. I hope it provides at least the same
API that tclodbc does without getting in the way too much.
What if I were to tell you that tclodbc is one of the model interfaces
being studied? It doesn't get everything quite right, but it comes
very close. Resurrecting it might indeed, as you and Tom Poindexter
observe, be a useful first step.
I'm a little nervous about using it for the long haul without substantial rework, partly because of your observations that it has substantial toolchain problems. Even in 2007, I find that C++ code frequently does. For whatever reason, it simply isn't as easy to get
deployment right for C++ as it is for C. (I salute the people who
manage to do so, and fully concede that they are either smarter
or more industrious than I am.)
--
73 de ke9tv/2, Kevin
.
- Follow-Ups:
- Re: Standard DBI Proposal
- From: Tom Poindexter
- Re: Standard DBI Proposal
- References:
- Standard Database Interface?
- From: tcltkdev
- Re: Standard Database Interface?
- From: Sean Woods
- Re: Standard Database Interface?
- From: tcltkdev
- Re: Standard Database Interface?
- From: Sean Woods
- Re: Standard Database Interface?
- From: thelfter@xxxxxxxxx
- Re: Standard Database Interface?
- From: Sean Woods
- Standard DBI Proposal (was: Re: Standard Database Interface?)
- From: Donal K. Fellows
- Re: Standard DBI Proposal (was: Re: Standard Database Interface?)
- From: Sean Woods
- Re: Standard DBI Proposal
- From: Donal K. Fellows
- Re: Standard DBI Proposal
- From: Darren New
- Re: Standard DBI Proposal
- From: Donal K. Fellows
- Re: Standard DBI Proposal
- From: tom.rmadilo
- Re: Standard DBI Proposal
- From: Donal K. Fellows
- Re: Standard DBI Proposal
- From: tunity5
- Re: Standard DBI Proposal
- From: Kevin Kenny
- Re: Standard DBI Proposal
- From: tunity5
- Standard Database Interface?
- Prev by Date: Swig as starkit anywhere ?
- Next by Date: Re: Swig as starkit anywhere ?
- Previous by thread: Re: Standard DBI Proposal
- Next by thread: Re: Standard DBI Proposal
- Index(es):
Relevant Pages
|