Composite index required does not exist, yet defined in index.yaml - google-cloud-platform

Question

I have some IoT devices that are sending some data into a Google Cloud Datastore.
The Datastore is setup as Cloud Firestore in Datastore mode.
Each row has the following fields:
Name/ID
current_temperature
data
device_id
event
gc_pub_sub_id
published_at
target_temperature
And these are all under the ParticleEvent kind.
I wish to run the following query; select current_temperature, target_temperature from ParticleEvent where device_id = ‘abc123’ order by published_at desc.
I get the below error when I try to run that query:
GQL query error: Your Datastore does not have the composite index (developer-supplied) required for this query.
So I setup an index.yaml file with the following contents:
indexes:
- kind: ParticleEvent
properties:
- name: data
- name: device_id
- name: published_at
direction: desc
- kind: ParticleEvent
properties:
- name: current_temperature
- name: target_temperature
- name: device_id
- name: published_at
direction: desc
I used the gcloud tool to send this successfully up to the datastore and I can see both indexes in the indexes tab.
However I still get the above error when I try to run the query.
What do I need to add/change to my indexes to get this query to work?

Answer 1

Though in the comment I simply suggest select * (that's the best way, I do think)
There is a way make your query work.
- kind: ParticleEvent
properties:
- name: device_id
- name: published_at
direction: desc
- name: current_temperature
- name: target_temperature
The reason why is select is done at the end and thus you need the index of current_temperature and target_temperature in a lower level.
Why I don't suggest this way is because, when your data grows and you need more combination of indexing just because of select specific columns. Your index size will grow exponentially.
But let's say if you sure you will just use this once and always query the data like this, then feel free to indexing it.
Or, if the connection bandwidth between your computer and google cloud is very small such that downloading more data causes you lag.