New to this card. I see the doc saying there are wildcard filters where you could generate the charts/bars dynamically?
Did I understand it correctly that I can do a “add all” card, where the charts/bars would grow when I add devices/entities?
Or maybe I’m way off…?
As to wildcards
The entity entry field accepts the
*wildcard and can automatically add multiple entities that match the provided pattern. Some examples:person.* - Add all entities from the person domain *door* - Add all entities that contain the term ‘door’ in the name, regardless of domain sensor.*door* - Add all entities that contain the term ‘door’ in the name, but only from the sensor domain * - Add all available entities in the list
Not 100% sure what you mean by the “add all” query. You can certainly interactively add as many entities as you want to a card, subject to processing power.
It really is a great card. I have one on my Lovelace test panel, and when I need to see some graphs, maybe as part of debugging stuff, I just delete the existing entities and add the stuff I want to look at.
I’m trying to see if this wildcard thing being a way to generate a “everything” view… basically replicate what “History” used to be, so that whenever any new entity is added to my system, sometime down the road, those entities would show up in the same “everything” view automatically.
… or if at least a way to manually identify and add new entities.
(… or here’s to hope that this feature request would get through. Yup I have voted.)
Reason being, as the codes around various integration evolves over time, entities would be added or retired or replaced or presented in a different way (say, different unit, or going from on/off to 0/1/2/3, etc.).
Hope this makes sense.
And if anyone have better ways to handle the situation, I’m all ears.
Look forward to you trying * as the entity id 
Ay!! Guess I should have read the docs more closely.
So the wildcard goes here:
… and not here, … I got nothing last night:
Right. The wildcards are supported by the entry field used to add entities over the UI (dynamically added entities). The entities added through the YAML do not accept wildcards (yet), you have to specify them individually.
It’s usually better to only add entities you are interested in anyway, rather than everything. This was a pretty bad design choice by the old HA history panel. While adding everything is possible, it will result in subpar performance and a lot of visual clutter. It’s best to use the card the way @nickrout mentioned, by adding what you need on demand. Other than the original history panel, the card can also create multiple independent history panels / views, showing a different set of entities. So for certain configurations, that can be a good setup too.
That said, I do plan on adding wildcard and filter support to YAML entities in the next release. This will allow the automatic display of entities matching the filter criteria. Including everything, if people really want to do that. I would still advise against it, because of the reasons mentioned above. But it will be possible.
Love it!! And thank you!
Totally understand that doing so could mean clutters and performance impacts. I do that simply because I want to figure out all those changes that weren’t obvious.
Hi.
I’m trying to use combineSameUnits: true in one of my multiple trends but it seems it only work for a variable.
Let me explain with an example
I have two variables pool_temp = piscina_temp with same unit (ºC) and another variable pool_ph = piscina_ph with the same unit (pH).
As you can see only appear in the same graph the “_temp” and there are two different graphs for the “_pH”.
The problem remains including a third variable (pool_orp and piscina_orp) with the same unit but different graphs 
Apparently, the combineSameUnits: true only works for the first variable. Doesn’t?
Combining works sequentially. The card will never second guess the user and will never change the order of graphs on its own. This is a fundamental design decision. And that includes combining. If you want all temperatures to combine in a single graph, then add them one after the other. If you add one temp, then one pH and again one temp, things will not combine because they’re not the same unit and the card will not change the order that you added them.
Offtopic, it seems you have a lot of gaps in your data. You might want to add the showUnavailable: false option to hide them.
Thank you very much, It works!
To improve:
Is it possible to rever displayed data upside down, as negatives? E.g. to show upload data above zero axis and download data below (though the sensor has positive values).
Not natively, but you could create a template sensor that would flip the sign of the download data, making it negative. Graphing this together with the unchanged upload data would give you the desired result.
That’s what I thought too 
though I do not like to create sensors if unnecessary… Already have almost 1000 in my system
You know what, open a feature request on github for a generic scale multiplier on line graphs. I had this in mind for awhile now, but keep forgetting about it. It could be used to flip your data on the fly (scale -1) or to do more generic unit conversions too.
V 1.0.25 is out !
Happy to say that the history explorer card is now fully localized and available in three new languages besides English: French, German and Spanish (thanks @Pharizna for the last one !). If your language is not yet supported and you’d like to see it added, your help is welcome !
Besides this, a new scale options was added to both YAML entities and dynamically added ones. It can be used change the scale of line graphs on the fly, without having to create new template entities. For example, you can use this to do unit conversions or sync the scales of multiple entities into a single graph. You can also use the scale to flip a graph upside down if you make it negative (@mirekmal).
Just installed and very happy with the result!
Speedtest results before:
Aand after (download flipped and scaled /10):
Thanks alot !!!
Nice !
I did notice that besides flipping the sign, it looks like the download speed was divided by 10, so the scale looks something like -0.1. Did you do this on purpose ?
Yes! In my case disproportion between maximum download (500 mbs) and upload (30 mbps) makes a bit hard to see variations on upload,if not scaled up. It is just very tiny vs download. So rescalling download makes it to look more like ‘percentage’ of available bandwidth. Obviously this could be achieved by creation of additional template sensors that would do similar recalculation, but this unneceserily multiply entities.
