 |
 |
Archives of the TeradataForum
Message Posted: Mon, 02 Feb 2009 @ 22:13:09 GMT
| Subj: | | Re: Data acces from JDBC, ODBC and CLI |
| |
| From: | | McCall, Glenn David |
| | How does data access from JDBC, ODBC and CLI compare (or differ) to each other for Teradata. The Comparison should include performance of
thedrivers in terms of time, memory required. Also, how does the performance load on Teradata vary when the clients use any of these
interfaces? | |
| | Our client is in a dilemma which one to use. Any help will be appreciated | |
I would suggest starting from a different perspective. First, what is your language of choice?
If your language choice is Java, then you are in the JDBC realm. Obviously you could use the others from Java if you really wanted to do so
(via JNI), but would you really want to?
If your choice is C, C++, or any of the 3-4GL languages, then you are most likely in the OBDC, OLEDB or .net provider. If you are talking C,
C++ or another tool that can make 32 bit API calls (e.g. VB) you could consider CLI.
In making the choice I would suggest looking at the features and complexity of each option. CLI is by far the most complex, but has the highest
performance (which may not be much of a factor for large long running queries).
If performance is important to you, then the best thing to do is measure the performance using a typical query that you use run against your
data in your environment. A program (or tool) that runs a query multiple times and measures the elapsed time is fairly trivial using each of the
technologies.
I hope this helps
Glenn Mc
| |
