Hi,

I loaded data from a file inthe table created in ICE and then reloaded the same table. As there are no primary keys in the ICE tables, I was able to duplicate the data in the table which seems to me something wrong. Please advise.

Also if I cannot do Insert, Delete or update on the Infobright table, how can I use it as a replacement of an Oracle database in my web application where I am doing transactions with the user.

Thanks,

Shahid

Shahid, please visit Infobright.com and download our IEE (Infobright Enterprise Edition) for a free 14 day trial. IEE has full DML capability,

Thanks,

Paul

Hi !

If you load the same data twice, you have them duplicated. I am not sure what makes you think there is something wrong with it - it is like with anything else e.g. if one puts a book on a shelf, and later one puts another copy of the book, then there are two books there.

If you need transactions, ICE is not for you (try IEE as Paul has mentioned already). Generally speaking, in order to help you with ICE, it would be good to know what was your motivation to try it as a replacement.

Hi Jansuz,

I understand that if I load the data twice, I will have data duplicated but what I am asking that if I cannot define a primary key for a table created in ICE, then how doess that data integrity maintained. My question arises from the use of conventional DBMS where one can specify a primary key and a new row with the same key is not allowed to insert.

There are two motivations for using the ICE:
1 - One for my masters thesis project to use Infobright database for using in the Rough Set Data Analysis algorithm ( a classification and decision ruloe generation technique).
2 - To evaluate Infobright to replace the legacy MSSQL database at my work.

Please clear.

Thanks,

Hi Shahid,

Infobright was designed mainly as a data warehouse engine. There is always a tradeoff between on-load functionalities (like integrity checking, constraints checking, key generation) and load performance scalability. Infobright was supposed to optimize the latter, i.e. to allow fast load of gigabytes of data. And this loading speed is the same for the first and the 100-th gigabyte. Thus, any constraints checking were “delegated” to a user’s ETL process which prepares data for loading.
Being primarily a data warehouse solution, Infobright should not be used “at any cost” instead other DBMS. In particular, typical OLTP usage pattern (many small insterts, deletes, updates), even if possible in IEE, is not what Infobright may outperform transactional DBMSs. The best usage pattern for Infobright is a seires of bulk loads, analytical ad-hoc queries, incidental updates/deletes (IEE).

Regards,

Hi !

To add a bit to Jakub’s comments: neither ICE nor IEE provide data integrity features, so primary keys are not available. The usual usage pattern is that data is loaded into ICE data warehouse from a standard DBMS (where integrity constraints are met), or through an ETL process. Therefore it is assumed that data loaded into ICE does not need integrity checks any more.

For your 1. aim, ICE can be a good choice, letting you to load quickly a lot of data and query them with reasonable speed.

Hi thanks for your posting people. This is a very useful posting about the data loading problem in ICE. From your posting I have got some reply also which I needed. I was also facing something like this. So I will try out those suggestions which you have got. Thanks.

ICE is used as a data warehouse where you only need it for FAST search.  I would not recommend ICE for any form of transactional database.  The benefit of having ICE is that you can also load InnoDB engine on your MySQL server (after installing ICE) as your transactional table engine because of its row locking vs. table locking ability. 

http://www.innodb.com/products/innodb_plugin/download/

If you want to load your data from a MySQL table into a BRIGHTHOUSE table, just simply dump the table in a csv file and then load it into the BRIGHTHOUSE table.