add a feature like Redis hget hset

[morteza102030]

It would be great if you could add a feature like Redis hget hset to lmdb

[kriszyp]

Like this?

function hget(db, id, property) {
  return db.get(id)?.[property];
}
function hset(db, id, property, value) {
  let record = db.get(id);
  if (record && typeof record === 'object') record[property] = value;
  else record = { [property]: value };
  db.put(id, record);
}

[morteza102030]

thank you for your answer

i store something like this in redis

r.hset('chat1', 'sararh', 5436)
r.hset('chat1', 'jack', 7396)
r.hset('chat1', 'bill', 5551)
...

chat1 have over 100000 users

r.hset('chat2', 'bill', 5456)
....
chat2 have over 100000 users
and the more

I think your example almost that i want

But I think this method is not suitable when our data has a large number of values (users)

I need and in Redis I can update thousands of users every second
It doesn't seem logical to load and save several hundred thousand users for each update
It is probably better if this process is done by C+ in the background

[kriszyp]

Ah, I see. I think you would probably want to use sub-databases for that:

chat1 = db.openDB('chat1')
chat1.put('sararh', 5436)

(and you should be able do many thousands of puts per second).

[morteza102030]

Because I have thousands of chats i don't know it's not problem i openDB all of them or not

Now when we use

let db = db.openDB('my-index', {
	dupSort: true,
	encoding: 'ordered-binary',
});

With this feature, we can easily and quickly have similar features redis sadd, srem, scard, sismember, smembers
This is a great options

If we had an option like this to create a simple nested keys like redis hset on one db it would be very interesting

[kriszyp]

Ah, ok, you can also use multi-value keys like this:

db.put(['chat1', 'sararh'], 5436);
db.put(['chat1', 'jack'], 7396)
db.get(['chat1', 'sararh']);

Each of these will be separate entries in the database (so db.get('chat1') won't return everything in chat1, but you can do db.getRange({start: 'chat1', end: ['chat1', orderedBinary.MAXIMUM_KEY]}) to iterate through all the entries in chat1).
This works as long as you are doing the default key encoding (ordered-binary), and you don't need dupSort for this, nor any special (value) encoding..