Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

How to store array of objects in Redis?

I have an array of Objects that I want to store in Redis. I can break up the array part and store them as objects but I am not getting how I can get somethings like

{0} : {"foo" :"bar", "qux" : "doe"}, {1} : {"name" "Saras", "age" : 23} 

and then search the db based on name and get the requested key back. I need something like this. but can't come close to getting it right.

incr id //correct
(integer) 3
get id //correct
"3"
SADD id {"name" : "Saras"} //wrong 
SADD myset {"name" : "Saras"} //correct
(integer) 1

First is getting this part right.

Second is somehow getting the key from the value i.e.

if name==="Saras"  
then key=1

Which I find tough. Or I can store it directly as array of objects and use a simple for loop.

 for (var i = 0; i < userCache.users.length; i++) {
            if (userCache.users[i].userId == userId && userCache.users[i].deviceId == deviceId) {
                return i;
            }
        }

Kindly suggest which route is best with some implementation?

like image 978
Saras Arya Avatar asked Aug 03 '16 15:08

Saras Arya


2 Answers

The thing I found working was storing the key as a unique identifier and stringifying the whole object while storing the data and applying JSON.parse while extracting it.

Example code:

client
    .setAsync(obj.deviceId.toString(), JSON.stringify(obj))
    .then((doc) => {
        return client.getAsync(obj.deviceId.toString());
    })
    .then((doc) => {
        return JSON.parse(doc);
    }).catch((err) => {
        return err;
    });

Though stringifying and then parsing it back is a computationally heavy operation and will block the Node.js server if the size of JSON becomes large. I am probably ready to take a hit for lesser complexity because I know my JSON wouldn't be huge, but that needs to be kept in mind while going for this approach.

like image 147
Saras Arya Avatar answered Oct 01 '22 21:10

Saras Arya


Redis is pretty simple key-value storage. Yes, there are other data structures like sets, but it has VERY limited query capabilities. For example, if you want to get find data by name, then you would have to to something like that:

SET Name "serialized data of object"

SET Name2 "serialized data of object2"

SET Name3 "serialized data of object3"

then:

GET Name

would return data.

Of course this means that you can't store two entries with the same names.

You can do limited text matching on keys using: http://redis.io/commands/scan

To summarize: I think you should use other tool for complex queries.

like image 38
my-nick Avatar answered Oct 01 '22 22:10

my-nick