nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Field Notes: Practical Tips From Professional Services

Eric Laurin

Featuring: Eric Laurin is a Principal Solution Architect and Community Captain

In Professional Services, new challenges often uncover practical lessons that can benefit the broader Board Community. The Board Platform offers many powerful capabilities, but some helpful options may not be immediately obvious. In this edition of Field Notes, I’m sharing a few practical tips that can help you avoid common issues and work more efficiently.

Here are a few tips for some cases that might fly under the radar:

When you make structural changes to a data model and then restore that data model from a backup, be sure to reload the data model into memory. This can be done by restarting the Board service or by going to Data Models > [data model name] > Summary > Unload from Memory. After the data model is selected again, it will reload into memory.

This step is especially important if you are using functionality that depends on the latest data model structure, including:

Subscription Hub’s Sync to Model, if configured
Calculations in data flows that use newly added structures

Reloading the data model helps ensure Board is working from the current structure and can prevent unexpected behavior after a restoration.

Another useful tip is for situations where you want to round a value down. For example, if you have a value such as 50.88 and want the result to be 50, you can use the INT(x) formula in a data flow. This works similarly to the ROUNDDOWN or TRUNCATE options you may be familiar with in Excel.
It is also helpful to keep cube type behavior in mind. If you calculate a value in a cube type such as Double or Single and then load it into an Integer cube, Board will apply behavior similar to rounding. For example, a value of 50.88 may round up to 51 when loaded into an Integer cube.

These small details can make a meaningful difference when configuring, restoring, or troubleshooting data models. Keeping them in mind can help you avoid confusion and maintain more predictable results in Board.

Have you run into this or a similar scenario in your workspace? Share your experience in the comments below so others can learn from it, too.

You might also like:
Value Realization with Kristine Young
New Process for Accessing Board Software Installers
Board's Latest Release Advances Continuous Planning
Like this content? Follow Eric for more from the author!

Find more posts tagged with

Data Modeling

Data Model Maintenance

Subscription Hub

Comments

DavideCorbetta

hi, while the problem you can have when you move from double or single to integer cube is easy to understand, I encountered more issues when combining single- and double-precision cubes.
The solution of simply using double-precision cubes for everything turned out to be unsuitable due to size and performance constraints.
Any suggestions?

Eric Laurin

Hello Davide,

My suggestion is that I generally don't use single type cubes at all, and leverage double for anything larger than an integer type. The reason for this is that the possibility of losing some detail behind the value and the calculation performance when converting from single to double or vice-versa overrode any concerns about size/performance of cubes. I have not seen an instance where I faced negative performance of a cube that was a double instead of single. Hopefully that helps!

Edgars Kancans

@Eric Laurin Thank you for sharing these field notes. I agree that unloading the database from memory before restoring is a good practice.

That said, I wonder if this step could be automatically triggered as part of the restore process. In most cases, the restore only takes a few seconds, so adding this as a Board Cloud triggered step likely would not add much wait time. More importantly, it would help ensure we are always working with a synced data model after a restore and in addition it saves us time when we train new Board developers.

Do you think this would be a good recommendation to submit to the Board Ideas Exchange?

DavideCorbetta

Hi, I'm not sure that's such instruction inside a procedure is a good Idea. that happen if the instruction is not the last one?

Hamza Mesbahi

@Edgars Kancans , makes sense to me. If a particular action is an expected part of a recurring process, such as before a restore, it likely makes sense to include it as an option in the procedure rather than relying on manual input each time.

Eric Laurin

@Edgars Kancans I always think in general if we have an idea we should submit it and generate the discussion.

Björn Reuber

Hello Eric,
to be honest, i do not agree with this kind of approach. From my point of view it depends on the usecase if double or single precission is necessary or not.
For example if there is a simple checkbox I would recommend to use single (ok in B10 it might have been integer), cause otherwise you will just double the size in memory and on disk for this cube.

Keep in mind, that double cubes also need double amount of data in memory and on disk

Regards

Björn