Hi

Probably you have to redesign the database schema. Using 1200 columns is rather uncommon for a single relational table.

Insertion of a single tuple in such a table with AUTOCOMMIT enabled, calls for >1200 IO requests against your disk to ensure transaction durability.

regards, Martin

On 15/09/15 14:20, Ying Zhang wrote:

...

On Aug 17, 2015, at 10:25, Vijay Krishna vijayakrishna55@gmail.com wrote:

Hi,

I was analysing the Insert performance of MonetDB...

I am trying to optimize a process where in a single tuple has to be inserted into a table of 1200 columns, which takes about 7.5 seconds presently... The query has values for only 5 columns out of 1200, yet the speed is still the same.. Is it possible to reduce this time even??

The columns are kept synced. Even if you only insert values to 5 columns, we’ll need to update all 1200 columns.

...

While digging on this, I came across this thread..

https://www.monetdb.org/pipermail/users-list/2005-August/000139.html

What is the insert speed relatively to other DBs?

Using SQL, with the current version we won't compete very well. We get much more acceptable speeds in 'raw' mode, where lots of the parsing overhead is removed.

Can anyone please help me understand what is the 'raw' mode that is explained in this thread??

He probably meant the COPY BINARY INTO feature: https://www.monetdb.org/Documentation/Cookbooks/SQLrecipes/BinaryBulkLoad

Regards,

Jennie

...

I am using Oct-2014 version on a CentOS machine with 128 GB RAM...

Thanks & Regards,

Vijayakrishna.P. Mobile : (+91) 9500402305. _______________________________________________ developers-list mailing list developers-list@monetdb.org https://www.monetdb.org/mailman/listinfo/developers-list

---

users-list mailing list users-list@monetdb.org https://www.monetdb.org/mailman/listinfo/users-list