topic Re: D3 Sankey Extension Node Duplicates in QlikView

D3 Sankey Extension Node Duplicates

Mon, 26 Jan 2026 16:26:21 GMT

Hello,

I am using the D3 inspired Sankey extension for Qlikview (https://github.com/brianwmunz/d3sankey-QV11)

I am able to get the sankey diagram to work, however I am seeing an issue with node replication. Below is an example of the issue I am facing.

As you can see, there are two nodes for Home -> Wallet. Ideally, all the Home -> Wallet records should go to the same node, with a smaller subset flowing to Register Wallet. Is this an issue with the Extension, or perhaps how I formatted the data? I have included the sample data file used to highlight the issue.

Thanks for your help!

Re: D3 Sankey Extension Node Duplicates

Tue, 03 May 2016 11:42:15 GMT

Qual dos graficos voce esperava?

Re: D3 Sankey Extension Node Duplicates

Wed, 05 Oct 2016 13:35:36 GMT

Hi Sorry for the delay, thank you for the response. I have tried your suggestion before, however ideally I would want the unlabeled nodes to be eliminated. Any idea where in the code I can update that?

Thank You!

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Thu, 06 Oct 2016 10:24:46 GMT

You should be able to have intermediaries as per Mike's example but I am not sure that the extension is handling this quite right.

If I get a chance I'll digest the code and see if I can come up with a better way of working with the nodes

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Thu, 06 Oct 2016 11:17:18 GMT

Bear with me.... just need to figure out the qlik data bit

Re: D3 Sankey Extension Node Duplicates

Thu, 06 Oct 2016 13:26:15 GMT

Ive begun looking at the code as well! But I am not a Jquery / Javascript user, so I look forward to what you come up with!

thank you!

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Thu, 06 Oct 2016 14:42:42 GMT

Is this the sort of thing you were hoping for? (needs some tidying up before being released)

Current bugs:

1- its backwards

2- can't handle orphaned data (e.g. a source but no target)

3- small bug in the way the array is built which causes a loop with certain data

Re: D3 Sankey Extension Node Duplicates

Thu, 06 Oct 2016 14:51:56 GMT

Yes!! That would be exactly what I am looking for. I am happy to Pad the data, ie add blank records for paths that are shorter than the longest path. I was trying to pad the paths, then just not add those nodes to the sankey. Not sure if that is a good / feasible option though and I havent gotten very far.

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Fri, 07 Oct 2016 06:14:53 GMT

Same here, I had never looked at JS / JQ until a couple of months ago!

I've slightly rearranged it so it takes Source, Destination and Size as the arguments which means you need to change the data source a bit, but it makes more sense this way.

I'll get the rough one up ASAP

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Fri, 07 Oct 2016 10:51:31 GMT

Nearly there, just trying to sort a click event to link it to data and then I think i'll be done

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Fri, 07 Oct 2016 13:50:57 GMT

BETA VERSION

This still needs some polishing and I need to better work on the Qlik code to convert a path to source and destination, been distracted with other work today unfortunately

Demo dashboard included

Extension expects: Source, Destination, Size

Double click event is enabled to select a node, it will select it as source if it is on the left or in the middle, or destination if it is the last item in the path.

You can also drag the nodes around to rearrange them

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Mon, 10 Oct 2016 08:43:00 GMT

Slightly updated qvw to turn a path into source / destination.

Re: D3 Sankey Extension Node Duplicates

Mon, 10 Oct 2016 12:56:32 GMT

Looking good! So this no longer takes in a delimited path?

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Tue, 11 Oct 2016 06:19:54 GMT

Hi Adam,

The extension itself doesn't, but the QVW has a script in it which will break down a comma delimited path into the requisite fields.

The QVW has two versions of the Sankey one running off the path data and one running off the source / destination data and both return the same values.

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Tue, 11 Oct 2016 12:50:23 GMT

Hi Adam,

Final update, just slightly tweaked it and made a checkbox for adding the number to the label.

All changes will now go via GIT (the new qar is on there)

https://github.com/adamdavi3s/QlikviewSankeyImproved

Re: D3 Sankey Extension Node Duplicates

Tue, 11 Oct 2016 13:44:23 GMT

Awesome! So I see there are still "dead" nodes, such as the -1 node on your git image. Also is there any way that the strokes or paths can take up the entire node?

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Wed, 12 Oct 2016 06:15:49 GMT

The dead one is there to show that it can handle something with no destination, I toyed with taking that out but actually I think it is important to show as it points you to a data issue.

How do you mean take up the whole node? As in the path should be the width of the node regardless of size?

Re: D3 Sankey Extension Node Duplicates

Thu, 13 Oct 2016 15:22:16 GMT

Is it a data issue though? If you think of this as clickstream data, they could click on the home page and not click on anything else. Just a thought! Thanks again for all of your help on this!

Re: D3 Sankey Extension Node Duplicates

adamdavi3s — Fri, 14 Oct 2016 06:29:53 GMT

Hmmm that is true but maybe not a valid use of a Sankey diagram in that case, how would you expect it to show those items that don't have any flow?

I guess the three possibilities would be:

1- as it is show a null when there is no destination

2- force something in there, e.g. a 'no onward path' node

3- entirely suppress items with no path

And not a problem, its good fun!

Re: D3 Sankey Extension Node Duplicates

Anonymous — Tue, 10 Apr 2018 21:28:17 GMT

Hi Adam,

My requirement is to show Successor and Predecessor for a Process. We have around 800 processes (jobs) which start after successful completion of Successor jobs and then Predecessor jobs start.

We have a straight table which display the list of Processes run in a day.

Process Name	Start Date	End Date
P1	4/9/18 12:10 PM	4/9/18 12:12 PM
P2	4/9/18 12:30 PM	4/9/18 1:30 PM
P3	4/9/18 2:30 PM	4/9/18 2:35 PM

The requirement is when to click on a process Sankey diagram should come up with the selected process as Process and its Successor and Predecessor

Could you please provide some help on this, any suggestions. Can i use your extension to add three nodes, when I double click the node it should become the middle node and its respective predecessor and successor shuld displau