I have developed a financial projection model for a business plan and the tables are working fantastically. But, now I have come to charting and the charts reset themselves if I change any parameter. And then spit out errors like this:
undefined is not an object (evaluating ‘e.columns().all().find((e=>e.summarySourceCol()===o.targetColRef)).id’)
Flow:
Add summary chart to my main table.
Switch to split series.
Select split series column. - Sometimes triggers x-axis and series to be reset.
Select x-axis - Same reset possibility and changes the split series.
One or both of these removes the data fields and filters.
The error above sometimes triggers and then the chart duplicates.
I would love to just create the summary tables I need and target a chart at it. In that scenario, I would just tell it which groups to use?
For now, I am going to have to move forward with minimal to no graphing.
Some complexities about my model: I am pre generating a maximum number of rows based on the number of months to project and the number of metrics I might track. Right now that’s 60 months and 75 metrics. That is 4500 rows. Currently this is using 32 of those metrics leaving a lot of “empty” rows.
Everything in this main table is a reference or reference list to another table or a fixed choice. Reference → month to know what month it belongs to. Reference → category…etc. The most complicated is a reference to “scenario” which defines the inputs to the calculation for the row.
I did find a workaround! Since my logic and other stuff is so complex, I got the idea of mirroring the source table.
To do that, I just copied the id’s and created references to every row in another table. In that table, formulas just pull in the column and I set the correct reference.
Now charts behave sanely. No goofy resetting, removing data, changing types, as long as I summarize against my “normalized” table.
It would be great to determine what I am doing that breaks the charting, but this solution definitely works. so I am going forward with it for now.
Glad you were able to find a workaround. There is a bug with Charts right now that is causing the issue you described. Our dev team is currently looking into it to get it resolved.
If you’re looking to do even more with Charts, you could use our Advanced Charts custom widget. It allows way more customizations, including changing colors! Learn more here: Custom - Grist Help Center
Another thing that I find frustrating is I can’t simply link a chart to a summary table? Or, at least, I can’t figure out how to do that. They don’t appear in the chart table lists.
This means I can’t do custom aggregations unless I also build my own table and use lookuprecords to summarize the data myself. Something as simple as switching to an average isn’t possible.
I feel like I have to be missing something here.
Thanks for reminding me about the advanced charts! I am going to try it.
Dear @Derrek, AFAIK all graphs are based on summary table. If you show the source data of the graph, you should see the name of the summary table, which you can open under “Raw Data”. There, you can edit the summary table and add all the calculations you need. The new columns can then be used in your graph, e.g. to create a new series in a circular graph.
DUDE! GENIUS! I should have made that connection when I first realized the raw data had all that stuff.
Other issue is a bunch of the Advanced chart things just do not work. Like legend groups. And selecting colors just stops for reasons I can’t determine.