472,096 Members | 1,184 Online
Bytes | Software Development & Data Engineering Community
Post +

Home Posts Topics Members FAQ

Join Bytes to post your question to a community of 472,096 software developers and data experts.

mySQL C API vector binding support?

Hello,
I have a "slight" problem with the mySQL C API. I want to bind vectors
for output/input prepared statements. However, the C API does NOT
support such data binding. Is there ANY other way to do a
bulk-INSERT/DELETE/UPDATE or bulk-SELECT in mySQL C API? I mean, one
can do it with a loop, but obiously the goal is to have a higher
performance.I tried it with a plain-old loop, but it is still slow.
Oracle supports such feature and it is considerably faster than the
usual loop-Insert (appr. 100 times faster for big queries) ! ! !
Please, advise.

Thanx.

Regards,

Dragomir Stanchev

Sep 25 '06 #1
1 4716
The|Godfather wrote:
Hello,
I have a "slight" problem with the mySQL C API. I want to bind vectors
for output/input prepared statements. However, the C API does NOT
support such data binding. Is there ANY other way to do a
bulk-INSERT/DELETE/UPDATE or bulk-SELECT in mySQL C API? I mean, one
can do it with a loop, but obiously the goal is to have a higher
performance.I tried it with a plain-old loop, but it is still slow.
Oracle supports such feature and it is considerably faster than the
usual loop-Insert (appr. 100 times faster for big queries) ! ! !
Please, advise.

Thanx.

Regards,

Dragomir Stanchev
Hi,
I seemed to found a partial solution to MY problem. However it works
only with INSERT.

When you want to add to your mySQL database many vectorS of elements or
other array-style things using prepared quires under mySQL C API you
can do it in 2 ways:
--->
The following tests were conducted using an array of 3000 vectors ,
each holding 100 elements of type int.
--->
1. Insert elements one by one. Obivous solution, but VERY
slowwww. For 300 000 elements (integers) it takes the staggering ~38
minutes doing INSERT to a remote ( not local) server.

2. The solution I used: just parse the Prepared query and instead
of:
"INSERT INTO tableName values()"
DO
"INSERT INTO tableName values(),(),(),().......(n) " ,

where n is vector<>.size() or whatever. Note that usually it is advised
to use a vector of size 100 elements cause it is the most efficient.
THE efficiency gain: ~15 seconds total time for the same 300 000
queries and the same remote mySQL server.
Just out of curiosity I testet the Oracle OCI integrated array
insertion mechanism and it turns out that Oracle was slower : ~19
seconds for 300 000 elements , 100 elements per one bulk insert.

Note: I tried to optimize it for UPDATE or/and DELETE but no success so
far. Any suggestions would be welcome.

Cheers,
Dragomir Stanchev

Oct 17 '06 #2

This discussion thread is closed

Replies have been disabled for this discussion.

Similar topics

1 post views Thread by Scott Bates | last post: by
5 posts views Thread by sandy | last post: by
reply views Thread by Mike Chirico | last post: by
9 posts views Thread by Chris Roth | last post: by
5 posts views Thread by star-italia | last post: by
6 posts views Thread by Tomasz J | last post: by
reply views Thread by leo001 | last post: by

By using Bytes.com and it's services, you agree to our Privacy Policy and Terms of Use.

To disable or enable advertisements and analytics tracking please visit the manage ads & tracking page.