KNX node for node-red

[sinn3r]

Hi sinn3r
what do you see, if you issue this from a terminal window:
locale

root@ioBroker:~# locale
LANG=de_DE.UTF-8
LANGUAGE=
LC_CTYPE="de_DE.UTF-8"
LC_NUMERIC="de_DE.UTF-8"
LC_TIME="de_DE.UTF-8"
LC_COLLATE="de_DE.UTF-8"
LC_MONETARY="de_DE.UTF-8"
LC_MESSAGES="de_DE.UTF-8"
LC_PAPER="de_DE.UTF-8"
LC_NAME="de_DE.UTF-8"
LC_ADDRESS="de_DE.UTF-8"
LC_TELEPHONE="de_DE.UTF-8"
LC_MEASUREMENT="de_DE.UTF-8"
LC_IDENTIFICATION="de_DE.UTF-8"
LC_ALL=

[TheMax74]

Hi sinn3r
TEST-MAIN and TEST-SUB are groups, not group addresses. The main and subgroup are never imported, because they're not group addresses.

(TEST-MAIN -> TEST-SUB) REAL GA # 1.001

This is correct, because TEST-MAIN is a main group, TEST-SUB is the subgroup and REAL GA is the group address.
Or maybe i've not understood what you're asking for?

[sinn3r]

Hi sinn3r
TEST-MAIN and TEST-SUB are groups, not group addresses. The main and subgroup are never imported, because they're not group addresses.

This is correct, because TEST-MAIN is a main group, TEST-SUB is the subgroup and REAL GA is the group address.
Or maybe i've not understood what you're asking for?

Yes, that's exactly what I meant. But you log this warning for every group line when importing the CSV.

[TheMax74]

Yupp... I'll take a look.

[TheMax74]

root@ioBroker:~# locale
LANG=de_DE.UTF-8
LANGUAGE=
LC_CTYPE="de_DE.UTF-8"
LC_NUMERIC="de_DE.UTF-8"
LC_TIME="de_DE.UTF-8"
LC_COLLATE="de_DE.UTF-8"
LC_MONETARY="de_DE.UTF-8"
LC_MESSAGES="de_DE.UTF-8"
LC_PAPER="de_DE.UTF-8"
LC_NAME="de_DE.UTF-8"
LC_ADDRESS="de_DE.UTF-8"
LC_TELEPHONE="de_DE.UTF-8"
LC_MEASUREMENT="de_DE.UTF-8"
LC_IDENTIFICATION="de_DE.UTF-8"
LC_ALL=

Please try to upgrade to the last node version.
I've 3 machines. 2 are on node v13 and 1 on v10.
The v.10 machine shows "AM/PM", the others not.

[sinn3r]

Please try to upgrade to the last node version.

Ah, very interesting!
As I run node-red in ioBroker and ioBroker seems to want Node v10 I will just stick to that time formatting till I can update Node to v13.

[TheMax74]

Yes, that's exactly what I meant. But you log this warning for every group line when importing the CSV.

Can you please paste here your exported CSV file?
I've tested another time, and all is fine for me.

NOT NEEDED ANYMORE

[TheMax74]

And there is another strange thing: When importing the CSV from ETS you give a warning "[warn] knxUltimate: WARNING: Datapoint not fully set (there is only the first part on the left of the '.'). I applied a default .001, but please set the datapoint with ETS" for each of the main/middle/sub group. When checking the imported group addresses it seems that these groups are not imported at all, which is for sure the right way to handle it, but the warning might confuse a few people.

Hello,
i've tested again. This warning warns you that the subtype is missing. For example, you only selected a boolean datapoint DPT1, but without subtype. The complete datapoint should be DPT1-1.
This is not a problem. The datapoint is automatically filled with the subtype 1. If you wanna fix this, you have to select the complete datapoint in ETS. I need to warn the user however.

The other part you asked for (skip if no datapoint) is ready to be published.

[sinn3r]

I've tested another time, and all is fine for me.

Oh I am so sorry, I totally misinterpreted the warning and found the problem. In the CSV there are some GAs with a datapoint of "DPT-1", but in ETS they are 1.001! This seems to be some kind of problem with the CSV Export. I will change them in ETS and see if this fixes the warnings.

UPDATE Hmm, no. I switched the datapoint in ETS to something else and then returned it to 1.001, after the new CSV export they are still "DPT-1". But as everything works as it is, there is no need to change it. Maybe you could write the name of the GA in the warning message. Currently you cannot see which GA is affected.

[TheMax74]

Yes exactly, see #173.
You should set the full datapoint in ETS before export.

V.1.1.55 is out with the "skip group address" option.

[sinn3r]

TheMax74
Very nice, works as expected

I finally found out, what causes the CSV export problem.
It's not enough to set the datatype of the GA, you also have to have a correct datapoint in the communication object of the linked device.
I have quite a few old GIRA presence detectors that have the output set to 1.* instead of 1.001.
After fixing them up I can not import the CSV without any warning or error.

Thanks again!

[martiko]

TheMax74
Is there a way to suppress the (in my case many) Debug-Messages druing import, caused by the faked datepoints?

During testing I need to deploy often and my debug window always gets "spammed" with these messages ;-)

[TheMax74]

Hello Martiko,
no, sorry. You need to live with that :-)
Setting the right datapoint is essential in any installation. Yes, this is a bit annoying, but indispensable.
I've had tooooo many users telling me "this doesn't work, that doesn't work", only because of lack of correct datapoints. I don't wanna to lose again so many time to fix the other's fault, without even a "thank you".
So, even you're an expert, please try to understand that if i put an option to disable the warnings, i'll be flooded by sentences like "this doesn't work, why you don't tell me that my datapoints are not properly set?".

PS: Instead of "full deploy", you can try to do a "modified node only" deploy.

[TheMax74]

TheMax74
Very nice, works as expected

I finally found out, what causes the CSV export problem.
It's not enough to set the datatype of the GA, you also have to have a correct datapoint in the communication object of the linked device.
I have quite a few old GIRA presence detectors that have the output set to 1.* instead of 1.001.
After fixing them up I can not import the CSV without any warning or error.

Thanks again!

Yes sinn3r, please see the Troubleshoot here, section "I'VE IMPORTED ETS FILE BUT THERE ARE WARNINGS ABOUT DATAPOINTS"

[martiko]

TheMax74
I read the current power consumption of a Device (Fritz!Dect200) and want so sent it so a KNX-GA. The consumption is measured in Watt but as I did not find a suitable DPT for Watt, I used DPT 9.024 so the currency is kWatt.
So I thought that I could do the scaling direclty in your KNX node using the conditional formating and selcting "Devide / 1000" (in German it's "Teilen / 1000"). But this had no influence on the output, the current Watt (e.g. 1.71) is sent unchanged to the KNX-Bus as before (without this setting). Have I misundertood this option, or have I misconfigured it, or do you have annother idea, why it is not working as expected?

PS: Question to all: Is there really no DPT with Currency Watt? For the normal usage, Watt would be more suitable (e.g. 1.71 vs. 0.00171)
PPS: I found DTP 14.056 for Watt, but I am not sure if a 4 Byte value makes sense, as I surely don't need this precision... Any suggestions?