streamly-lmdb

Stream data to or from LMDB databases using the Haskell streamly library.

Requirements

Install LMDB on your system:

• Debian Linux: sudo apt-get install liblmdb-dev.
• macOS: brew install lmdb.

Quick start

{-# LANGUAGE OverloadedStrings #-}

module Main where

import Data.Function ((&))
import qualified Streamly.Data.Fold as F
import qualified Streamly.Data.Stream.Prelude as S
import Streamly.External.LMDB
  ( Limits (mapSize),
    WriteOptions (writeTransactionSize),
    defaultLimits,
    defaultReadOptions,
    defaultWriteOptions,
    getDatabase,
    openEnvironment,
    readLMDB,
    tebibyte,
    writeLMDB,
  )

main :: IO ()
main = do
  -- Open an environment. There should already exist a file or
  -- directory at the given path. (Empty for a new environment.)
  env <-
    openEnvironment "/path/to/lmdb-database" $
      defaultLimits {mapSize = tebibyte}

  -- Get the main database.
  -- Note: It is common practice with LMDB to create the database
  -- once and reuse it for the remainder of the program’s execution.
  db <- getDatabase env Nothing

  -- Stream key-value pairs into the database.
  let fold' = writeLMDB db defaultWriteOptions {writeTransactionSize = 1}
  let writeStream = S.fromList [("baz", "a"), ("foo", "b"), ("bar", "c")]
  _ <- S.fold fold' writeStream

  -- Stream key-value pairs out of the
  -- database, printing them along the way.
  -- Output:
  --     ("bar","c")
  --     ("baz","a")
  --     ("foo","b")
  let unfold' = readLMDB db Nothing defaultReadOptions
  let readStream = S.unfold unfold' undefined
  S.mapM print readStream
    & S.fold F.drain

Benchmarks

See bench/README.md. Summary (with rough figures from our machine^†):

• Reading:

  • For iterating through a fully cached LMDB database, this library has roughly a 110 ns/pair overhead compared to C. (Plain Haskell IO code has roughly a 70 ns/pair overhead compared to C. The two preceding figures being similar fulfills the promise of streamly and stream fusion.)
  • By using unsafeReadLMDB instead of readLMDB, we can get the overhead down to roughly 100 ns/pair.
  • By additionally using the readUnsafeFFI option (to use unsafe FFI calls under the hood), we can get the overhead down to roughly 40 ns/pair.

• Writing:

  • For writing to an LMDB database, this library has roughly a 210 ns/pair overhead compared to C. (Plain Haskell IO code has roughly a 100 ns/pair overhead compared to C. The two preceding figures being similar fulfills the promise of streamly and stream fusion.)
  • By using the writeUnsafeFFI option (to use unsafe FFI calls under the hood), we can get the overhead down to roughly 140 ns/pair.

• For most Haskell programs, these differences will not cause problems. (For instance, note that merely opening and reading 1 byte from a file with C already takes us tens of microseconds.)

^† May 2023; Linode; Debian 11, Dedicated 32GB: 16 CPU, 640GB SSD storage, 32GB RAM.