Structuring a query response with PostgreSQL - node.js

Question

I am trying to construct a query to return data from multiple tables and build them into a single array of objects to return to the client. I have two tables, incidents and sources. Each source has an incident_id that corresponds to an incident in the first table.
Since there can be more than one source I want to query for the incidents, then on each incident add a key of source that has a value of the array of associated sources. The desired final structure is this:
{
"incident_id": 1,
"id": "wa-olympia-1",
"city": "Olympia",
"state": "Washington",
"lat": 47.0417,
"long": -122.896,
"title": "Police respond to broken windows with excessive force",
"desc": "Footage shows a few individuals break off from a protest to smash City Hall windows. Protesters shout at vandals to stop.\n\nPolice then arrive. They arrest multiple individuals near the City Hall windows, including one individual who appeared to approach the vandals in an effort to defuse the situation.\n\nPolice fire tear gas and riot rounds at protesters during the arrests. Protesters become agitated.\n\nAfter police walk arrestee away, protesters continue to shout at police. Police respond with a second bout of tear gas and riot rounds.\n\nA racial slur can be heard shouted, although it is unsure who is shouting.",
"date": "2020-05-31T05:00:00.000Z",
"src": ['http://google.com']
}
Here is the route as it stands:
router.get('/showallincidents', (req, res) => {
Incidents.getAllIncidents()
.then((response) => {
const incidents = response.map((incident) => {
const sources = Incidents.createSourcesArray(incident.incident_id);
return {
...incident,
src: sources,
};
});
res.json(incidents);
})
.catch((err) => {
res.status(500).json({ message: 'Request Error' });
});
});
Here are the models I currently have:
async function getAllIncidents() {
return await db('incidents');
}
async function createSourcesArray(incident_id) {
const sources = await db('sources')
.select('*')
.where('sources.incident_id', incident_id);
return sources;
}
When this endpoint is hit I get a "too many connections" error. Please advise.

Answer 1

I found a solution. What I decided to do was to query the two tables independently. Then I looped through the first result array, and within that loop looped through the second array, checking for the foreign key they share, then when I found a match, added those results to an array on the original object, then returned a new array with the objects of the first with the associated data from the second. Models are unchanged, here is the updated route.
router.get('/showallincidents', async(req, res) => {
try {
const incidents = await Incidents.getAllIncidents();
const sources = await Incidents.getAllSources();
const responseArray = []
// Reconstructs the incident object with it's sources to send to front end
incidents.forEach((incident) => {
incident['src'] = []
sources.forEach(source => {
if (source.incident_id === incident.incident_id) {
incident.src.push(source)
}
})
responseArray.push(incident)
})
res.json(responseArray)
} catch (e) {
res.status(500).json({
message: 'Request Error'
});
}
});

Answer 2

Are the two tables in the same database? If that is the case, it is much more efficent to do the primary/foreign key match by doing an SQL join. What you have implemented is "nested loop join" which might not be the optimal way to match depending on the value distribution of the primary key. You can search SQL join algorithms to see examples and pro/cons.
If the tables are in different databases then indeed client side join is likely your only option. Though again if you know something about the underlying distribution it might be better to do a hash join