State: PUT vs POST

If for example i have a LightSwitch on OpenHab, i can switch it “ON” by a PUT (directly to the /state URI) or by a POST (to the item URI)

When is it preferred use PUT or when is it preferred use POST ?
If is there no difference i may use POST for all items, also for the items that have a /state ?


If you do a PUT, then it is the same as a postUpdate(), while a POST is the same as a postCommand().

So, if you want to actually send a command, then you need to use POST.


Let me explain better my question (sorry for my english):
I can switch on a lamp whit a PUT:
by sending a ON payload.
But i can also switch on the same lamp whit a POST:

For this example (not for other kind of items) is there a problem by using
I can completely remove the PUT from my client code and use only POST
command, also for change the state of the lamps, or is there a special
condition where i need PUT ?

It seems to me that we never really need the PUT verbs.

Thanks for your answer
Il 17/ago/2015 20:01, “Chris Jackson” ha

As above, the two should do different things. Have you tested this with the demo setup or are you actually controlling REAL hardware?

thanks, maybe now i “see” :smile:
I tested it only with demo setup, so i cannot view the difference, now i understand that postUpdate() change only the state of the items, so example on a Classic UI i view the lightswitch change it position but the real Lamp connected to it will not switched on.

But this means also that at the end of a postCommand() automatically is called, always, a postUpdate() ?
(i think this because on the demo setup sending a POST to a Lamp by a javascript page, also change the switch position on ClassicUI)

P.S.: maybe you mean sendCommand() ? because there is no postCommand() in the Action documentation

1 Like

Correct - an update is made automatically with OH after a command is sent. From memory there is an option to disable this but it don’t see it specified in the wiki.

1 Like

Sorry. My phone posted before I was finished…

You are probably right about action names - I was looking in the source code where postUpdate and postCommand are the methods that are called, but in the action documentation I think your naming is correct.


1 Like

So this is a very important info, because example if we open a Garage whit a POST, and have a limit switch installed, we want that the /state is changed at the end of movement, when the limit switch is activated, not immediately after the POST command.

Do you know where is possible search this info, or the name for search on github code ?

Ok - this took me a while to find and it doesn’t appear to be in the wiki!

This link shows an example of how to make a pushbutton. It uses a (seemingly undocumented) command in the item definition called autoupdate="false", and it looks like this is the way to stop the status of the item automatically updating when the command has been sent. It’s not something I’ve personally used, so you might want to give it a test :smile:

I think if you want to operate your garage door as you suggest, then you will need to add this option. However, I would suggest that in your example you actually need two items - one that is the actuator that is linked to the command to open the door (and therefore shows the state of the motor that is moving the door) and the other the status of the sensor that shows if the door is fully open or not…


1 Like

Many bindings are directional, such as the http and mqtt bindings, so it will depend on how your device is interfaced with openhab. In fact you can even mix and match bindings within item definitions.
For example, a device that is connected that is controlled/polled through http get requests (in this case a holding relay that needs to be turned on and off though a separate request, and a separate contact that shows the status). Notice the direction > for outbound or < for inbound:

Switch Switch1 "Switch 1" {http=">[ON:GET:] >[OFF:GET:] <[*))]"}
1 Like