If you are unable to create a new account, please email support@bspsoftware.com

 

News:

MetaManager - Administrative Tools for IBM Cognos
Pricing starting at $2,100
Download Now    Learn More

Main Menu

decimal datatype problem on enabling DQM

Started by kiran.timsina, 27 Mar 2013 09:56:19 AM

Previous topic - Next topic

kiran.timsina

I am trying to see the change in performance in my DMR package by enabling and disabling DQM. But I get an error related to datatype on enabling DQM    right at the beginning while creating a metadata. The error is: "XQE-DAT-0004: A decimal with precision 1 and scale 4 is invalid." My table column is numeric(38,4). Once I import it, FM assigns a precision of 0 and scale 4, which is obviously invalid. The column datatype is fine if DQM in unchecked. Any solutions? (I am using FM 10.1.1)

My whole point is to see performance difference on using DQM. Will there be any observable improvement in report run time on using DQM?

bdbits

The difference in performance is going to depend a lot on the data source/model. The most noticeable gains will be seen with a DMR package over a relational data source.

For your errors I am not sure. You might try just enabling DQM at publishing time. As I understand it, it will use the JDBC drivers and DQM engine if you do this. To be honest I am not sure what actually turning it on in the model does, other than probably checking for the stricter modeling rules required for successful use of DQM. But you may see some performance increase with just enabling during publishing.

We only have 1 DQM-based model here thus far and there were a few gotchas. Overall it seems to be alright. We have not tried converting an existing package to use DQM.

Good luck and post back the results if you see performance differences.

Rahul Ganguli

SQL generated by a DQM model is different in Syntax than CQM model.
Once you change from CQM to DQM there may be few hickups, but those get rectified easily.

Hav e done lot of migration to DQM, have recieved better results always and very good results in case of Multidimensional models.
BUt, There is one scenario where We were using Oracle as Datasource and relational model where we experienced negative results, don't know why and didn't get time to diagnose due to delivery pressure.

Regards,
Rahul

kiran.timsina

Hi Rahul,
Does it make any difference if I don't check DQM right at the beginning while creating a metadata? Will may package be DQM if I enable DQM only while publishing the package?

Thanks.

MFGF

Quote from: kiran.timsina on 02 Apr 2013 08:10:20 PM
Hi Rahul,
Does it make any difference if I don't check DQM right at the beginning while creating a metadata? Will may package be DQM if I enable DQM only while publishing the package?

Thanks.

It makes no difference to the published package if you select DQM when creating the project or when publishing the package. If you don't select DQM at the beginning, it simply means you have the option to publish the package using CQM or DQM.

Cheers!

MF.
Meep!