Space in IQ_System_Main DB Space in SAP PA 2.2 Expert mode

[Former Member]

Hi Everyone

I imported around 4,50,000 rows from CSV file and applied Auto Classification Algorithm over it. But when I run the model, below error code gets generated.

[Sybase][ODBC Driver][Sybase IQ]You have run out of space in IQ_SYSTEM_MAIN DBSpace.

-- (s_blockmap.cxx 4616)

Any idea how to resolve it?

regards

Ranajay

[PPaolo · Accepted solution]

Hello Ranajay,

  do you have other software on the same machine which could be using Sybase IQ as well behind the scenes?  I have seen a similar issue on a machine where there was a BusinessObjects server using SybaseIQ as repository. There might be a contention of Sybase resources.

Could you try to:

= reboot your machine and re-do the workflow from scratch?

= reproduce the problem with the same dataset but on another computer if possible?

Also, in general, please notice that we can try and help from this forum but if you have a serious problem impacting your business it would be better to contact the SAP customer support where the case can be followed properly and patches provided, if needed.

Thanks and regards,

PPaolo

[Former Member · Accepted solution]

Hi Pierpaolo

I restarted the machine and when I run the data again with the same model with same number of datasets, I am getting a different error now and it took around 3 minutes to process. Is it because of large datasets. See the below snapshot

[Former Member · Accepted solution]

I guess the reason behind this weird behavior was large dataset. Because what I did now was randomly selected 3 columns and and selected one target variable for Auto Classification Algorithm. It took about 1 min but the model run successfully. But earlier when I was selecting around 40 columns for 4,50,000 rows the above mentioned random errors were happening.

Is there any solution for this?

[Former Member · Accepted solution]

Hi Ranajay,

Thanks for reporting this issue... we ran some tests and discovered under certain circumstances, the underlying database file isn't able to expand to contain the larger datasets used for analysis.

A dynamic detection and extension mechanism has been now implemented and will be available with the next release due shortly.

A point to be noted here is that local desktop based analyses are bound by the processing power and memory of the desktop. For larger datasets and complex processing, PA should be used with SAP HANA.

Regards,

Jayant

[Former Member · Accepted solution]

Hi Jayanta

Thanks for handling this issue in your upcoming release. Though can you just quantify the amount data on which prediction can be done smoothly using the desktop based SAP PA. Will be helpful for others who might be facing such issue.

regards

Ranajay

[Former Member · Accepted solution]

Hi Ranajay,

There are a number of variables that influence the size of the dataset in memory that includes the number of components you have in a chain and kind of manipulations you are performing etc...

I ran an Auto Classification on a million records census dataset with 15 features and it worked fine. It failed with the error you experienced when I added a data type definition node to change the column types.

I am testing a few options to handle this via configuration and I'll post details soon.

Regards,

Jayant

[Former Member · Accepted solution]

Thanks Jayanta.

[achab · Accepted solution]

Hi all,

What is the current status on this one?

Thanks & regards

Antoine

[Former Member · Accepted solution]

I am not facing this issue anymore in SAP PA 2.3. Thanks everyone

.

Regards

Ranajay