Assess NodeJS storage solutions about sheikah HOT 5 CLOSED

The most interesting I have found is level, which bundles levelup, leveldown and encoding-down
https://www.npmjs.com/package/level

• level
  • exports a function that returns a levelup instance when invoked
  • bundles the current release of levelup and leveldown
  • leverages encodings using encoding-down
• levelup
  • Fast and simple storage.
  • A Node.js wrapper for abstract-leveldown
• leveldown
  • provides a pure C++ binding to LevelDB
• LevelDB
  • Is a simple key-value store built by Google
  • supports arbitrary byte arrays as both keys and values, singular get, put and delete operations, batched put and delete, bi-directional iterators and simple compression using the very fast Snappy algorithm.
• encoding-down
  • It allows you to specify an encoding to use for keys and values independently
  • The encoding is applied to all read and write operations
  • The default encoding is utf8 which ensures you'll always get back a string. You can also provide a custom encoding like bytewise - or your own

from sheikah.

After checking more storage solutions I think this one is better for our needs as it offers serialization:
https://www.npmjs.com/package/lowdb

lowdb

• Small JSON database for Node, Electron and the browser. Powered by Lodash
• Syntax

    db.get('posts')
       .push({ id: 1, title: 'lowdb is awesome'})
       .write()

• FileSync, FileAsync and LocalStorage accept custom serialize and deserialize functions. You can use them to add encryption logic

    const adapter = new FileSync('db.json', {
      serialize: (data) => encrypt(JSON.stringify(data))
      deserialize: (data) => JSON.parse(decrypt(data))
    })

from sheikah.

Even though I like the lowdb API the most, I have a couple of concerns about such solution:

• It seems to me that FileSync handles data as text and writes to the DB file in text mode, so an encoding like base64 would be needed to represent encrypted data and thus we are forced to bear a 33% expansion.
• Its robustness and consistency is questioned for non-trivial use cases, and it has no error correction at all (see discussion here).Their own README says it becomes unreliable as the DB file grows.

Regarding Level, it's true that being a pure key/value store, it lacks rich QL capabilities. But apart from that, it is much stronger than lowdb in the two areas I'm more concerned about:

• It supports binary mode by instatiating levelup with { encoding: 'binary' }. In addition, it has native compression (Google's Snappy).
• LevelDB is not totally free from corruption, but has been proven to be a reasonably solid solution for embedded DBs (i.e.: Bitcoin Core uses LevelDB for storing the whole blockchain).

@kronolynx What do you think?

from sheikah.

@aesedepece after reading the discussion and your arguments, I think level should be used.

from sheikah.

Great, we can consider this done. OK to close.

from sheikah.