Wildcards in csv definitions? Or how do I solve variations in ID?

[kstuken]

The control unit of my Ochsner GSWM 10 HK heatpump broke a few weeks ago, and the replacement control unit upgraded the device from OTE 2 plus to OTE 3 plus. As a result, a lot of the ebus messages changed and I had to update large parts of my definition files, which has worked well, so far.

In the process, however, a number of broadcast messages that delivered useful data have disappeared. (Probably because the device that provided a web interface for my heatpump is not compatible with OTE 3 plus and hence now defunct, and so the chatter between that device and the heatpump has stopped). After quite a bit of searching I have now found some ebus messages that seem to carry the data that I am after. But I struggle with creating a definition for them, as these messages are weird with respect to message ID.

To illustrate, here's three messages that I have identified that carry the data I am chasing currently (which is the setpoint temperature for the heating circuit). I have inserted some spaces for readability:

2025-10-24 06:26:48.945 [update notice] received unknown MS cmd: 1091 1031 0b 3f 82 80 0d 02 e8 03 00 00 5c 01 / 070101004a41007f  
2025-10-24 06:41:37.835 [update notice] received unknown MS cmd: 1091 1031 0b bd 82 80 0d 02 e8 03 00 00 5d 01 / 070004004ebf007f
2025-10-24 14:12:45.511 [update notice] received unknown MS cmd: 1091 1031 0b 6b 82 80 0d 02 e8 03 00 00 48 01 / 070004004e6d0000  

The data I am after is the last two bytes of the master payload, ie "5c01", "5d01", and "4801", respectively. (SIN/10)

But how do I capture these messages? I have identified the second and third byte of the master payload ("8280") as a reliable identifier. However, the first byte is actually a counter, running from 01 to ff for all 1091 messages (which are sent multiple times every minute). Hence, I am not sure how to specify a message ID.

I had hoped to be able to use a wildcard for the ID definition, eg something like "??8280" (see below), but that does not appear to work, I get an error in the log.

So my question is:

• is there some sort of wildcard I could use in this situation, or something similar, to define a more flexible ID that ignores the first byte and only takes the second and third byte
• if not, is there a different way to achieve what I want?

Your input / hints would be most appreciated.
Thank you!

type	circuit	name	comment	QQ	ZZ	PBSB	ID	field1	part	type	divider/values	unit	comment	field2	part	type	divider/values	unit	comment
r	21576	temperature.heating.flow.setpoint_2		10	91	1031	??8280	unknown	m	HEX:6					m	tempt

[kstuken]

Hi,

Can I try to bring this up once more, please?
In the csv definitions, is there a way to skip / ignore a byte?

I have a bunch of messages from the same source QQ, target ZZ, and primary and secondary byte PBSB.

The ID part of these messages it tricky, though.
The first byte seems to be a counter, continuously incrementing for each message with that same QQZZPBSB.
But following that counter byte, I can identify groups of messages based on the following bytes.

So I want to capture eg all messages with ID xx8280, where xx should be ignored.

How can I do that?

[john30]

you need to define defaults for e.g. read and add an ignored field there. then using that default can extend the id accordingly. here is an example of that:
ehp:22

ah wait a moment, you want to have a part of the id being ignored. thats not possible at the moment.
are you sure this is a counter rather than a checksume or similiar?

[kstuken]

Ah, shame. Thanks for responding and confirming, though.

Yes, pretty sure about the counter. Analyzing all 1091 1031 messages in my log, that byte is monotonously increasing by 1 for each message and then resetting at ff, irrespective of the remainder of the message.
Following that counter byte I can discern clear patterns that correspond to ID and payloads in the other messages.

As a workaround I will probably now just dump the payload of every 1091 1031 message into Home Assistant, and then use a python script to process the payload.

If at some point ebusd might become capable of handling this, that would be wonderful. I don't know the inner workings of the engine processing this, but being able to do wildcards like "xx0180" or "??0180" for ID would already solve my problem.