Best Of

Hi Pierre,

Sorry about that.  I really should not try to respond with a rushed how-to between work tasks.  I will correct my second response later however, attached is a capsule that should help you to understand the process required to create dates and values once you've entered a seed date.  The NEXEL is performed after being called in a procedure.  The values are then saved in 2 identical cubes but one is for DATES and the other for VALUES.  Hopefully it will make sense once you look at the attached model.

Instructions:

1.   Enter seed date in left hand cube.

2.   Select [Button].

3.   Verify that right hand cubes are as expected results.

Hi Jean-Paul Baar

The behaviour you describe can be avoided if you close and re-open the target screen. 

There is probably some layout related memory mapping 'thingy' going on which gets corrupted when you initially edit the target screen but then try to paste objects from other screens which leads to the behaviour you are witnessing.

The absolute method to copy multiple objects from one screen to another would look something like this:

Method to copy multiple objects and paste them in exactly the same place as their point of origin. 

1. Close the target screen, saving any changes you may already have made.
2. Open the source screen.
3. Open the target screen.
4. Copy the objects from the source screen.
5. Paste onto target screen.

Hope this helps 

Regards,

Paul

Hello Enric,

What a great question.  I had not thought to look at this behaviour and now that I have, I'm surprised at the outcome. 

So, the good news first.  There is no need to consider this an 'issue' as WEB behaviour is different to that of the (outgoing) client model.  Whilst there are options for the client, they do not relate to the web version as all entries, made in WEB, are written and calculated on once entered into a dataentry cell, regardless of client screen settings.  You can test this:

1. Open your CLIENT version
2. Deselect Save/Undo DataEntry
3. Deselect Refresh on DataEntry
4. Open a web instance of your model
5. Open the model in WEB.
6. Enter a new value into the data entry box on your web instance.
7. Do nothing else whatsoever in WEB
8. Refresh your client screen (F5).
9. Web values show in Client.

Now, my surprise is that 'all entries, made in WEB, are written and calculated on once entered into a dataentry cell'.  So, if you have dataentry functionality on a spreadsheet type matrix, your screen will refresh and re-calculate each and every time you change an entry.  this could make dataentry a very tedious operation as such 'refresh' behaviour is very inefficient, as well as preventing you from backtracking to reverse mistakes, as was possible on the client model.

Our own best practice (BP) on all screens which provide dataentry is to implement 'workflow' processes that capture @User previous settings to allow backup and restore functionality.  we also implement datestamping and scenario entity members where required.

Hope this helps

Hi Massimo,

I just caught your post and wanted to tell you that you can achieve what you need by copying (dataflow) the values to a dedicated cube with a single Entity.  The members of the entity are the columns you outline.  Once complete, you can use algorithms, manual selections and Cube selection (data entry) to determine what columns (Members of the entity) are shown.  You can even suppress NULL value columns, which cannot be done when building with blocks.

This approach is only suitable in instances like yours, where the values are of the same datatype (date, value, text, etc).  You can mix the values but everything must be put into a text cube and so you lose the ability to calculate or export to excel easily.

Check out Adam Elvins post from 05 Nov 2017 Conditional Display of Columns who discussed the exact same issue.  My own post is here ADHOC Report Builder and requests that BOARD Intl incorporate the ability to hide columns at will from outside of the object builder.

Please feel free to ask me any questions on building what you need.

Regards,

Paul

Hi Lukas Xuereb,

I'm feeling your pain my friend.  As an active community member for six months, you will be aware of the transition of BOARD from Desktop to Web, commencing with BOARD 10.5.  So, I can only imagine your post was in a fit of frustration as there is no chance of BOARD updating an outgoing product.  We are all, some more than others, mentally scarred by the BOARD design. 

There are many instances where the design becomes maddening.  Appreciating that the journey you are on is the same journey that alot of us have walked, you may already be aware that formatting dataviews and charts is an absolute LOL blast, requiring one to learn the location of each component and then where it is formatted; not all borders are formatted with the tick box HOWTO: Hide ALL Dataview border lines    and Chart Markers & Legend, ha, don't get me started.

So,  please post your woes, it's healthy to vent frustrations,  but know that you have support from veterans who wear their BOARD scars proudly and are willing to jump to the aid of fellow BOARDERS.  You are not alone.  BOARD didn't get it all right the first time...no, second...  OK, so BOARD are continually revising and improving their offering .  The Roadmap is aspirational and User feedback is greatly received but rarely acknowledged Review process for Idea : Feedback. 

But, on the plus side... it's only 62 days to Christmas...Yay, I love Christmas.

Hi Malav Shelat

You raise an interesting point regarding a loss of interest.  Aside of the community website issues, the user experience of BOARD from a developer perspective is becoming tedious where once it was fresh and exciting.  The user interface, overall style choices, feel of the models and lack of expected functionality are also having a negative impact on the opinions of clients and their belief that BOARD is the right solution.  

The website issues on their own would be ok if they were in isolation.  However, they are now mirroring the problems with those of the product.  For all the great strengths of BOARD (Data Entry, Nexel, ETL) the product appears to be losing its focus, along with the lack of maintenance on the community website. 

The quality and non-compatability issues are compounding other issues which are either out-performed by rival software or vital functions which do not follow established best practice modelling for intuitive use, such as:

• Map Object.   The map object which does not have any ability to carry out proximity analysis like tableau; with v10.6 it also now has bugs like the need to show the lat and long just to plot points, which then show also in the ugly, non-formattable, information pop-out. 
• Bullet Chart.   The terribly executed bullet chart.  Ironically, this data visualisation was specifically conceived for efficient dashboard data visualisations and is exceptional in Power-BI and easily created in Excel.
• GUI.   Confusing, un-intuitive and misleading graghical user interface decisions like:
  • PRESENTATIONS which are in actual fact custom reports and screens that users can use to default their own setting in.
  • Document Creation (PDF) Should be a simple and intuitive process to create executive level reports with user comments and storable configurations.
  • ICONS.   The iconography was never great but the use of an eye icon to indicate entity selections on or off is an example where BOARD appear to be introducing a new approach to interaction in place of a more established on/off or tick indicator.
  • NEXEL crossview calculations that do not work in the upgraded capsule.   Considering that nexel is used to conveniently replace complex SQL processes and that the calculations are directly used in the core function of analysis, this issue in itself raises questions regarding the BOARD process for testing and releasing features and upgrades.

Issues, like those above, may be as a result of releasing versions and features too early in order to meet deadlines, a lack of communication/control between the company and the programmers or an error in concept and design direction. The impact of these issues is that I'm beginning to enjoy rival products more as they now prove to deliver on client expectations by doing precisely what I want, easily and with great control and fidelity from start to finish.  Complimenting the rival software, community websites are established, populated with technical experts, as with the BOARD Community, polished and professional in feel and execution.

Regards,

Paul

I would like to add my support for both points from Etienne CAUSSE. 

I would also like to expand on his second one regarding adhoc reporting - I will probably submit an idea on it if someone has not already.The most prevalent issue I have with my clients is the lack of ability to create ahoc reports within the software.  I do not have the resources to train any client in how to use the Self-Service features. It would be great to have a 'lite' report creation tool, which is a simple drag-n-drop report creation tool for all users and based only on the objects currently on screen.

I think one of the issues is the way licencing is pitched and perceived. I am a developer and so have  a licence to develop, our data analysts have poweruser licences but our core user population of internal and external business users are Lite or LitePlus licence holders.  The issue here is that lite and liteplus users can also be 'power(ful)users' as they are also the ones taking our datamodels and augmenting them into their own highly skilled and professional roles.  It is they that display how good a BI tool is and they all want to slice and dice data in order to create an adhoc report.  And that is despite the fact that some of us can do some crazy creative stuff with BOARD; Want to set local security filters for specific users and back up the file  - go right ahead:

Here's the logic (based on an original idea by Robert Heatley - the workflow master)

There is no tool on the market that can do all of what BOARD can, in my opinion.  Data entry and variable handling used to build workflow processes without the need of a programming language is immensely powerful.  Here at Avison Young, our standard models also provide integrations with BOX.COM, Leverton AI systems, Google maps and other web sources and services.  But I can tell you that my external clients are like "Meh" when they see it because their levels of expectation in 2019 are such that what we do is just expected today.  The problems come when a, business educated, senior director cannot just get the report they want and asks for a datadump to excel so someone else can do it for them - This is not what they want to do at the conference on their IPAD, I have to mitigate this issue by supporting the analysts so that they can create reports for their clients; pulling the proverbial rabbit out of the hat.

If I could have one wish for a future BOARD feature, just one, then simple adhoc report creation for all users would be it. 

   ... set the data free .

This response does not help in resolving this issue, however...

I think a change of, previously understood, behaviour that has not been fully explained, which does not provide a choice and so results in unexpected behaviour where once the outcome was understood, constitutes an issue (wow, sorry about the sub-clauses #smiley-face#).  

From my point of view I cant tell if this is an issue or only a changed behaviour

Platform upgrades such as 10.3 are welcome but they should be implemented and documented so that  developers can also implement and document changes to the client.

I really love BOARD BI but feel that I am made to run uphill on each update.