Data on demand on property change

[Marko nidari]

I have the following case: One data connection with several tables. I add on table to dxp first and then addrows from other tables. Data is fetched on demand for each new table based on property. For example we have 10 nodes but on property value only 3 nodes willfetch data, other will return 0 rows. When property is changed, next 2nodes will return data while others will return 0 rows (including from those first 3 nodes).

Row count of the rows fetched is small (few thousands) while table has lots of columns (several hundreds). What we noticed is that on each property change RAM is consumed. For example on first load Spotfire consumes 1 GB addtional RAM and total RAM consumtion is for example 10 GB. On property change 1 GB is consumed and total consumption is 11 GB, and so on. Spotfire just adds GB and on property change no RAM is released. With lots of property changes sooner or later RAM is used and disk engaged and from that point on Spotfire is stuck. All the caching can be disabled - on data connection and withing data on demand data loading setting but the result is the same.

Question: why does Spotfire consumes RAM in a cumulative way on each property changeThe point of fetching data with on demand setting and property is effective way of fetching data to dxp to optimize data loading and RAM consumption. On property change no old data is needed anymore (when not cached). Is there a way how to release RAM in this case The only way I can do that now is to close the Spotfire session. Even closing the file (withouth closing the client) does not release total RAM used bythat dxp.

[David Boot-Olazabal]

Hi Marko,

I do not know the specific technical details when it comes to charging memory, but it is normal behavior of Spotfire when you load data it consumes a certain amount of memory. When you add data, through on-demand queries, this base amount of memory (let's assume it is 10Gb) will be extended with the 1Gb from the on-demand queries. And this will repeat itself, as long as your dashboard is open (which in fact keeps the base 10Gb in cache plus the added Gb's from the on-demand side).

You can solve this bylowering the number of columns (do you really need all those hundreds of columns) or adding more memory to your laptop (I do not know how much is needed and if that even is a possibility for you).

If you are interested in the more technical part of Spotfire capturing data loads/using memory, I would advise to create a support ticket via support.tibco.com.

Kind regards,

David