Multiple users modify values simultaneously in a Table Editor on server

Hi guys,

we have a scenario where multiple users access a KNIME dashboard on server and modify/enter some values on a certain column of a Table Editor. The users might access the dashboard at the same time, but modify different rows of the table (rule is that each user has its own product to modify).
The output of the workflow should be an excel file.

So far, we tried accessing the dashboard from two users and modify different rows. The excel would contain only the value modified by the last action.

Would this be possible to implement in KNIME, that the table acts as a repository, saving the changes from every user? Or at least lock the dashboard for other users somehow, when a user is using it?

Any recommendations are very welcomed. Thanks a lot in advance for your great work!

Regards,
Denis

1 Like

Hi Denis,

currently, I don’t think this is directly possible. Behind the scenes, each instance of the dashboard is a separate object (job) that doesn’t sync data with other jobs.

However, you could design the dashboard to only save a single column (user selects which column to save) and have another job in the background combine all individual columns into a single file.

With our upcoming version 4.5 we will also introduce an Excel Sheet Updater node, that only updates specific cells and leaves others untouched. That may be something worth looking into.

Collaboration is definitely an interesting use-case, and I will pass this on internally to look into.

Kind regards
Marvin

4 Likes

On another note:
Working on the same files/tables simultaneously is often best covered with transactions (i.e. working with databases). Maybe you (or your knime admin) can also discuss your use-case in more detail with one of our solution engineers. Please feel welcomed to contact your technical account manager to set things up.

Kind regards
Marvin

4 Likes

Hi Marvin,

thanks a lot for the answer and the hints. We will see what makes the most sense for us.

Best regards,
Denis

2 Likes

Hi again,

now that I am thinking it, there should be a not so elegant way of achieving this.

We produce a table (KNIME table or Excel) after each user’s job, saving every entry the user did. The table name would have a timestamp attached, just so that they don’t get overwritten. At the end of the day, we simply run another flow to merge this single tables into one master table and delete the single ones.

I would like to know your opinion on this.

Thanks & regards,
Denis

1 Like

Hi Denis,

yeah, I think that will work just fine. In current versions, the Excel Reader can read in all files in a single folder at once, or you can use the node List Files/Folders – KNIME Hub to iterate over files.
If you have redundant columns in the files you might be able to join on them. It may also suffice to simply concatenate the files, depending on how exactly your data looks.

In case confidentiality and time allow, you’d be welcome to share an example, maybe someone can use it in the future, or suggest improvements.

Still, in the long term this likely is a problem best solved without having Excel sheets as intermediate results.

Kind regards
Marvin

2 Likes

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.