Reservoir

No results found

Lean4 Postgresql

This library embeds PostgreSQL syntax and semantics into Lean 4. The rules for the type system are taken from the Postgres documentation and source code version 17.

Installation

Lake

Add the following dependency to your lakefile.lean :

import Lake
open System Lake DSL

require Postgres from git "https://github.com/FWuermse/lean-postgres" @ "master"

package conn where

@[default_target]
lean_exe conn where
  moreLinkArgs := #["-lpq", "-L/<your path to libpq>"]
  root := `Main

Usage

Requirements

Requires a running Postgres instance e.G. via Docker:

docker run -d --name postgres -e POSTGRES_PASSWORD=password -v ~/my-volume:/var/lib/postgresql/data -p 5432:5432 postgres

Code

import Postgres

open LibPQ Query QueryAST QuerySyntax

def stringTables (table : Option (List (List String))) : String :=
  match table with
  | none => "Error"
  | some t => "\n".intercalate (t.map (", ".intercalate .))

def schema : Schema := [("employee", [("id", "employee", .bigInt)]), ("customer", [("id", "customer", .bigInt), ("date", "customer", .date)])]

def queriesOnSchema : SQL Unit := do
  let query := pquery( schema |-
    SELECT customer.date AS d
    FROM employee, customer
    WHERE -(employee.id * -2) = customer.id
      AND customer.id != 0 ∶ [("d", "customer", .date)] )
  let res ← sendQuery query
  IO.println <| stringTables res
  IO.println query

def main : IO Unit := do
  let conn ← login "0.0.0.0" "5432" "postgres" "postgres" "password"
  queriesOnSchema.run {conn}
  testQueries.run {conn}
  pure ()

$ ./.lake/build/bin/query

    d      
2003-01-03
1997-11-29

Please find more examples in the example folder.

Supported DataTypes

DataType	Usage	Limitations
NULL	empty type
Integer	32 bit integer
BigInt	64 bit integer
bit	fixed length bit array
varBit	variable length bit array
boolean	TRUE or FALSE
char	fixed length character array
varchar	variable length character array
text	character array of arbitrary length
date	Date of the form yyyy-mm-dd	validation not month specific, no other formats supported
double	64 bit floating point number	only double no floats supported