How to save and return JavaScript object with Subarray in Normalized SQL
I use postgres-node
, but I think this is a problem for anyone who has javascript objects that have subarrays that they want to store in SQL. I have javascript objects with different number (any length) of array of functions:
{
name: "Ted",
features: ['Red Hair', 'Blue Eyes']
}
so when I have several of them, javascript formats it like this:
[
{
name: "Ted",
features: ['Red Hair', 'Blue Eyes']
},
{
name: "Ann",
features: ['Brown Hair', 'Blue Eyes', 'Big Smile']
}
]
It's great! But how do I get this back from the database after normalization? I have normalized this in my database like this:
people
Table
+---+------------+
|id | Name |
+---+------------+
| 1 | Ted |
| 2 | Ann |
+---+------------+
features
table
+---+--------------+
|id | feature_name |
+---+--------------+
| 1 | Red Hair |
| 2 | Blue Eyes |
| 3 | Brown Hair |
| 4 | Big Smile |
+---+--------------+
and people_features
connection table
+---+-----------+-------------+
|id | person_id | feature_id |
+---+-----------+-------------+
| 1 | 1 | 1 |
| 2 | 1 | 2 |
| 3 | 2 | 2 |
| 4 | 2 | 3 |
| 5 | 2 | 4 |
+---+-----------+-------------+
If I do something like this:
SELECT name, feature_name
FROM people
JOIN people_features ON people_features.person_id=people.id
JOIN features ON people_features.feature_id=features.id;
I am getting one row for each individual person. This is not what I want.
What I get:
[
{
name: "Ted",
feature_name: 'Red Hair'
},
{
name: "Ted",
feature_name: 'Blue Eyes'
},
{
name: "Ann",
feature_name: 'Blue Eyes'
},
{
name: "Ann",
feature_name: 'Brown Hair'
},
{
name: "Ann",
feature_name: 'Big Smile'
}
]
What I want:
[
{
name: "Ted",
features: ['Red Hair', 'Blue Eyes']
},
{
name: "Ann",
features: ['Brown Hair', 'Blue Eyes', 'Big Smile']
}
]
It sounds awful! Now I need to go through them and combine identical people into the object of one person. My other option seems to be making a request for people
SELECT id, name
FROM people;
What will be returned:
[
{
id: 1
name: "Ted"
},
{
id: 2
name: "Ann"
}
]
And then I need to loop through and make a separate SQL query for each individual?
For every person:
SELECT feature_name
FROM features
JOIN people_features ON features.id=people_features.feature_id
WHERE people_features.person_id = $1
($ 1 is the ID of the person I'm scrolling through)
And then I'll be back (for Ted):
[
{ feature_name: 'Red Hair' },
{ feature_name: 'Blue Eyes' }
]
Then I need to remove them from my objects (just get the string) and then add them to the object.
Is one of them the best way to do it? I feel that they are both very ineffective.
source to share
Wao Zong's answer is wonderful. For those who are interested, here is the minimum version of what I used in my node code to get it to work with node-postgres
:
var pg = require('pg');
var config = {
user: process.env.PG_USER || null, //env var: PGUSER
password: process.env.DATABASE_SECRET || null, //env var: PGPASSWORD
host: process.env.DATABASE_SERVER || 'localhost', // Server hosting the postgres database
port: process.env.DATABASE_PORT || 5432, //env var: PGPORT
database: process.env.DATABASE_NAME || 'lukeschlangen', //env var: PGDATABASE
max: 10, // max number of clients in the pool
idleTimeoutMillis: 30000, // how long a client is allowed to remain idle before being closed
};
var pool = new pg.Pool(config)
pool.connect(function (err, client, done) {
if (err) {
console.log('There was an error', err);
} else {
client.query(
'SELECT name, array_agg(feature_name) ' +
'FROM people ' +
'JOIN people_features ON people_features.person_id=people.id ' +
'JOIN features ON people_features.feature_id=features.id ' +
'GROUP BY people.id;',
function (err, results) {
done();
console.log(results.rows); // This was exactly the array I wanted
}
);
}
});
source to share