When data are published into EdgeX message bus, besides the actual device value, it also has some additional values, such as event created time, modified time etc in the event. Sometimes these values are required for data analysis. This article describes how to use functions provided by eKuiper to achieve the goal.
The data structure received from EdgeX message bus is list as in below. An Event
structure encapsulates related metadata (ID, DeviceName, ProfileName, SourceName, Origin, Tags), along with the actual data (in Readings
field) collected from device service.
Similar to Event
, Reading
also has some metadata (ID, DeviceName... etc).
If upgrading from eKuiper versions v1.2.0 and before which integrates with EdgeX v1, there will be some breaking changes of the meta datas due to the refactor of EdgeX v2.
Pushed
, Created
and Modified
for both events and readings are removed.Device
for both events and readings are renamed to DeviceName
.Name
of readings is renamed to ResourceName
.So how the EdgeX data are managed in eKuiper? Let's take an example.
As in below - firstly, user creates an EdgeX stream named events
with yellow color.
Secondly, one message is published to message bus as in below.
demo
with green colortemperature
& Humidity
with red color.metadata
that is not necessary to "visible", but it probably will be used during data analysis, such as DeviceName
field in Event
structure. eKuiper saves these values into message tuple named metadata, and user can get these values during analysis. Notice that, metadata name DeviceName
was renamed from Device
in EdgeX v2.Thirdly, a SQL is provided for data analysis. Please notice that,
events
in FROM clause is yellow color, which is a stream name defined in the 1st step.temperature
& humidity
are red color, which are the Name
field of readings.meta(deviceName)
in green color, which is ued for extracting DeviceName
field from Events
structure. The SQL statement will filter data that device names are not demo
.Below are some other samples that extract other metadata through meta
function.
meta(origin)
: 000Get 'Origin' metadata from Event structure
meta(temperature -> origin)
: 123Get 'origin' metadata from reading[0], key with 'temperature'
meta(humidity -> origin)
: 456Get 'origin' metadata from reading[1], key with 'humidity'
Please notice that if you want to extract metadata from readings, you need to use reading-name -> key
operator to access the value. In previous samples, temperature
& humidity
are reading-names
, and key
is the field names in readings.
However, if you want to get data from Events
, just need to specify the key directly. As the 1st sample in previous list.
The meta
function can also be used in SELECT
clause, below is another example. Please notice that if multiple meta
functions are used in SELECT
clause, you should use AS
to specify an alias name, otherwise, the data of previous fields will be overwritten.
SELECT temperature,humidity, meta(id) AS eid,meta(origin) AS eo, meta(temperature->id) AS tid, meta(temperature->origin) AS torigin, meta(Humidity->deviceName) AS hdevice, meta(Humidity->profileName) AS hprofile FROM demo WHERE meta(deviceName)="demo2"
meta
function can be used in eKuiper to access metadata values. Below lists all available keys for Event
and Reading
.