Integration with database using native connector

by Veselin Pizurica`

Hi everybody, in this post we will show how to turn any database into a waylay sensor.

Turning database into a sensor

Waylay comes with sandboxed nodejs VM in which scripts get executed, similar to AWS lambda. Sandboxed VM’s come with number of packages that are pre-installed. You can simply access them from the script, no need to declare them. Feel free to check the Waylay documentation in order to see the complete list. In this blog we will use native Microsoft SQL connector.


In waylay we don’t talk about connectors, but rather about sensors and actuators. In that respect, you can think of the database record as a sensor: e.g. Customer, or an actuator: e.g. database log entry (Incident etc). More about the philosophy behind our concepts you can find here blogpost. In this example, I will search for a device in the database and send back the record item as the raw data - if the record is found. I will also use states “Found”/“Not Found” to tell the user of this sensor whether the record has been found in the database. This can be handy in case you use the conditional execution of the sensors feature, where the outcome of one sensor can trigger execution of the next sensor video link. For instance, you may want to implement a use case where the smart meter collection triggers search in the ERP database, after which an e-mail needs to be send to the user of this meter (while the e-mail address is in the ERP database). In this example “Found” state of the ERP sensors would trigger e-mail actuator, and the e-mail address itself would come from the raw data of the same sensor. So, let’s see the code:

 find a customer by meter ID. Since we assume that we first get a meter data and only then search for a customer, assumption is that the
 customer ID needs to be lookup based on the meter sensor. Therefore we are looking at the meterID from the rawData. If the use case was other way around
 (first we fetch customer), then we would need first to get a customer and use meterId to find a meter measurements.
var server = options.globalSettings.DB_HOST;
var user = options.globalSettings.DB_USER;
var password = options.globalSettings.DB_PASSWORD;
var table = options.globalSettings.DB_TABLE;
var database = options.globalSettings.DB_DATABASE;

var meterNumber;
try {
    //this call always throughs the exception
    console.log("try to fetch meter from the raw data");
    meterNumber = waylayUtil.getRawData(options, options.requiredProperties.meterNumber);
} catch(err){
    console.log("try to fetch meter from the input argument");
    meterNumber = options.requiredProperties.meterNumber || waylayUtil.getResource(options);

if(meterNumber === undefined ){
    send(new Error("meter number not found"));
} else {
        var config = {
        user: user,
        password: password,
        server: server,
        database: database
        mssql.connect(config, function(err) {
            if(err) {
                console.log("error, enable to connect")
                 send(new Error(err));
            } else {
                var request = new mssql.Request();
                request.query('select * from ContactInfo where MeterNumber='+meterNumber, function(err, recordset) {
                if(err) {
                    send(new Error("error"));
                } else{
                        var res = {
                            observedState : "Not Found"
                        if(recordset.length == 1) {
                            res.observedState = "Found";
                            res.rawData = recordset[0];
                            send(null, res);
                        } else
                            send(null, res);
        console.log("Exception connection error:")
        send(new Error(err));

As you can see, here I used global settings to fetch database settings. I also used as the input argument a meterInfo, which can either come from the raw data of the other sensor, or as the simple input argument, or if that is missing, also it can come from the resource input (should you wish to use template input feature). Once you define your sensor, just make sure you define in the raw data all database columns that this sensor will return. Again, this is a good practise, it is not really needed, but if you want someone else to use this sensor, it is always good to tell what is the info that comes out of this sensor. More over, if you want to use autocompletion feature of the UI, you must define it anyway.

So I have many tables, does it mean I need to define all of them as sensors?

Yes indeed. If you design logical blocks in your system such as assets, users, customers, it makes sense to create all of them as sensors, since your business guy, as everyone else, will be able to visually understand what is going on. But, we also made platform to be flexible enough, and in the Belgium spirit, we also allow you do just pass as the input argument any query you like and return back any raw data you find interesting. If that is what you want to do, go ahead and create one generic DB sensor, and use node labels to describe what is going on.