Easy. Simple. Powerful. Generate (complex) SQL queries using magical Elixir SQL dust.

Every language has its commonly used libraries / gems / packages to interact with databases. Ruby has ActiveRecord and Elixir has Ecto. They provide a lot of functionality which are very useful but when it comes to quickly and easily querying tabular data they require too much hassle:

• you have to describe models representing tables (example)
• you have to describe how to join tables (example)
• using the query DSL requires a bit of reading and understanding how to use it

Actually, you do not want to waste time specifying how to join tables and thinking about table aliases when you have followed the standard naming convention. And you do not want to think about putting a condition in the WHERE or HAVING statement.

The solution is to think in paths (e.g. company.tags.name) and letting the package do the magic regarding joining table and to use SELECT statement aliases to determine HAVING statements.

Enter SqlDust. It makes it as easy and simple as possible for the developer to generate SQL queries:

• no models setup
• no joins specifications
• no DSL to learn

Just focus on what really matters! ;)

Based on standard naming conventions, SqlDust will determine how to join tables. You just have to specify from which resource (table) to query from and which columns to select using paths:

$ iex -S mix
iex(1)> SqlDust.from("users", %{select: ~w(id first_name company.category.name)})
{"SELECT\n  `u`.`id`,\n  `u`.`first_name`,\n  `company.category`.`name`\nFROM users `u`\nLEFT JOIN companies `company` ON `company`.`id` = `u`.`company_id`\nLEFT JOIN categories `company.category` ON `company.category`.`id` = `company`.`category_id`\n",
 []}
iex(2)> IO.puts elem(v, 0)
SELECT
  `u`.`id`,
  `u`.`first_name`,
  `company.category`.`name`
FROM users `u`
LEFT JOIN companies `company` ON `company`.`id` = `u`.`company_id`
LEFT JOIN categories `company.category` ON `company.category`.`id` = `company`.`category_id`

:ok
iex(3)>

Composable queries are also possible:

$ iex -S mix
iex(1)> import SqlDust.Query
nil
iex(2)> select("id, last_name, first_name") |> from("users") |> where(["company.id = ?", 1982]) |> to_sql
{"SELECT `u`.`id`, `u`.`last_name`, `u`.`first_name`\nFROM users `u`\nLEFT JOIN companies `company` ON `company`.`id` = `u`.`company_id`\nWHERE (`company`.`id` = ?)\n",
 [1982]}
iex(3)> IO.puts elem(v, 0)
SELECT `u`.`id`, `u`.`last_name`, `u`.`first_name`
FROM users `u`
LEFT JOIN companies `company` ON `company`.`id` = `u`.`company_id`
WHERE (`company`.`id` = ?)

:ok
iex(4)> "users" |> adapter(:postgres) |> to_sql
{"SELECT \"u\".*\nFROM users \"u\"\n", []}
iex(5)> IO.puts elem(v, 0)
SELECT "u".*
FROM users "u"

:ok
iex(6)>

Composable queries using Ecto models are also possible (using Ecto simple example):

iex(1)> import Ecto.SqlDust
nil
iex(2)> Weather |> to_sql
{"SELECT \"w\".*\nFROM weather \"w\"\n", []}
iex(3)> IO.puts elem(v, 0)
SELECT "w".*
FROM weather "w"

:ok
iex(4) City |> select("id, name, country.name") |> where(["country.name = ?", "United States"]) |> to_sql
{"SELECT \"c\".\"id\", \"c\".\"name\", \"country\".\"name\"\nFROM cities \"c\"\nLEFT JOIN countries \"country\" ON \"country\".\"id\" = \"c\".\"country_id\"\nWHERE (\"country\".\"name\" = ?)\n", ["United States"]}
iex(5)> IO.puts elem(v, 0)
SELECT "c"."id", "c"."name", "country"."name"
FROM cities "c"
LEFT JOIN countries "country" ON "country"."id" = "c"."country_id"
WHERE ("country"."name" = ?)

:ok
iex(6)>

To install SqlDust, please do the following:

1. Add sql_dust to your list of dependencies in mix.exs:

   def deps do [{:sql_dust, "~> 0.3.10"}] end

2. Ensure sql_dust is started before your application:

   def application do [applications: [:sql_dust]] end

Generating SQL queries has never been simpler. Just invoke the SqlDust.from/3 function. It accepts the following arguments:

• resource (required) - Usually this is the table from which you want to query from
• options (required) - A map containing info about what the query should contain (e.g. :select, :where, :group_by)
• schema (optional) - A map containing info which overrule the defacto derived schema

options = %{
  select: "id, name, COUNT(orders.id) AS order_count, GROUP_CONCAT(DISTINCT tags.name) AS tags, foo.tags",
  group_by: "id",
  where: ["name LIKE '%Paul%'", "order_count > 5", "foo.tags = 1"],
  order_by: "COUNT(DISTINCT tags.id) DESC",
  limit: 5
}

schema = %{
  customers: %{
    tags: %{
      cardinality: :has_and_belongs_to_many
    }
  }
}

SqlDust.from("customers", options, schema) |> elem(0) |> IO.puts

"""
SELECT
  `c`.`id`,
  `c`.`name`,
  COUNT(`orders`.`id`) AS `order_count`,
  GROUP_CONCAT(DISTINCT `tags`.`name`) AS `tags`,
  `foo`.`tags`
FROM customers `c`
LEFT JOIN orders `orders` ON `orders`.`customer_id` = `c`.`id`
LEFT JOIN customers_tags `tags_bridge_table` ON `tags_bridge_table`.`customer_id` = `c`.`id`
LEFT JOIN tags `tags` ON `tags`.`id` = `tags_bridge_table`.`tag_id`
LEFT JOIN foos `foo` ON `foo`.`id` = `c`.`foo_id`
WHERE (`c`.`name` LIKE '%Paul%') AND (`foo`.`tags` = 1)
GROUP BY `c`.`id`
HAVING (`order_count` > 5)
ORDER BY COUNT(DISTINCT `tags`.`id`) DESC
LIMIT ?
"""

As of version 0.1.0, it is to possible compose queries (thanks to Justin Workman for the request):

import SqlDust.Query

select("id, last_name, first_name")
  |> from("users")
  |> where("company.id = 1982")
  |> where("last_name LIKE '%Engel%'")
  |> order_by(["last_name", "first_name"])
  |> to_sql
  |> elem(0)
  |> IO.puts

"""
SELECT `u`.`id`, `u`.`last_name`, `u`.`first_name`
FROM users `u`
LEFT JOIN companies `company` ON `company`.`id` = `u`.`company_id`
WHERE (`company`.`id` = 1982) AND (`u`.`last_name` LIKE '%Engel%')
ORDER BY `u`.`last_name`, `u`.`first_name`
"""

As of version 0.1.1, it is to possible compose queries using Ecto(!) models:

import Ecto.SqlDust

City
  |> select("id, name, country.name, local_weather.temp_lo, local_weather.temp_hi")
  |> where("local_weather.wdate = '2015-09-12'")
  |> to_sql
  |> elem(0)
  |> IO.puts

"""
SELECT
  "c"."id",
  "c"."name",
  "country"."name",
  "local_weather"."temp_lo",
  "local_weather"."temp_hi"
FROM cities "c"
LEFT JOIN countries "country" ON "country"."id" = "c"."country_id"
LEFT JOIN weather "local_weather" ON "local_weather"."city_id" = "c"."id"
WHERE ("local_weather"."wdate" = '2015-09-12')
"""

At default, SqlDust generates queries with MySQL quotations except for when using Ecto models because then it defaults to Postgres. You can specify the adapter using the adapter function after having piped the Ecto model:

import Ecto.SqlDust

City
  |> select("id, name")
  |> adapter(:mysql)
  |> to_sql
  |> elem(0)
  |> IO.puts

"""
SELECT `c`.`id`, `c`.`name`
FROM cities `c`
"""

SqlDust should automatically determine the correct database adapter of the Ecto model of course. So that will be added in the following release.

As of version 0.2.0, when generating the SQL query, SqlDust returns a tuple containing the SQL query string along with a list containing values which the database should interpolate safely.

Enjoy using SqlDust! ^^

As of version 0.3.0, SqlDust returns a tuple with an additional list of variable keys which corresponds to the passed options[:variables].

To have it clear:

• SqlDust will return a tuple of two elements when NOT having passed variables
• SqlDust will return a tuple of three elements when having passed variables

Run the following command for testing:

mix test

All the SqlDust features are tested in test/sql_dust_test.exs, test/sql_dust/query_test.exs and test/ecto/sql_dust_test.exs.

• Query from the database

• Automatically derive database adapter using Ecto.Repo
• Support has through associations
• Support polymorphic associations
• Add additional documentation to the README
• Add doc tests for internal functions

Copyright (c) 2017 Paul Engel, released under the MIT License

http://github.com/archan937 – http://twitter.com/archan937 – [email protected]

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.