SQL Tools

A rust crate meant to make SQL queries simple and communication between various SQL versions easy. The goal is to have most major variations of SQL compatible, however only Oracle SQL and SQLite are available right now. This is an evolution of oracle_sql_tools. This project is constantly changing, check out the change logs to see what's new.

How To Use

In your cargo.toml file:

[dependencies]

sql_tools = "<CURRENT_VERSION>"

# chrono is required if you're working with dates 

chrono = "0.4.41" 

To start using SQL Tools, you need a new connection.

Oracle SQL

use sql_tools::sql_variations::OracleConnect;

let username = "new_user";
let password = "password123!";
let connection_string = "my-secure.connection/database";

let conn = OracleConnect::new(connection_string, username, password)?;

SQLite

You can open a SQLite connection in a db file or open a connection in memory.

use sql_tools::sql_variations::SQLiteConnect;

// Open connection from a file
let path = "<PATH_TO_DB_FILE>";
let conn = SQLiteConnect::from_path(path);

// Open connection from memory
let conn = SQLiteConnect::in_memory();

Once you established a connection type, you can use the various methods in this crate to interact with your database. These options are select, update, insert, create, delete, and alter. The data types that are supported by default can be found in the docs under the ToSQLData trait. You can implement ToSQLData for your own enum or struct to make integration into your application easy.

SQLDataTypes

This is the enum that is used to apply the proper type to the data that's being selected, updated, or inserted.

pub enum SQLDataTypes {
    Varchar(String),
    Number(i64),
    Float(f64),
    Date(chrono::NaiveDateTime), 
    NULL,
}

SELECT

For the select method, you add the table you want to select from, then you can choose specific columns, functions, varchars, or everything from a particular table. To see these options more in-depth, look at the Column enum. You can add a conjunction statement to filter out the rows you want, just like a SQL query.

let product_id = ColumnProps { name: "product_id".to_string(), table: "regional_sales".to_string() };
let state = ColumnProps { name: "state".to_string(), table: "national_sales".to_string() };
let city = ColumnProps { name: "city".to_string(), table: "regional_sales".to_string() };
let product_ids = WhereArg::Values(vec![1001.to_sql_fmt(), 4567.to_sql_fmt()]);
let state = WhereArg::Like("Tex%");
let cities = WhereArg::Values(vec![SQLDataTypes::Varchar("Austin"), SQLDataTypes::Varchar("Dallas")]);
let columns = vec![
    Column::Value(ColumnProps { name: "revenue as national_rev".to_string(), table: "national_sales".to_string() }),
    Column::Value(ColumnProps { name: "revenue as regional_rev".to_string(), table: "regional_sales".to_string() }),
    Column::Value(ColumnProps { name: "city".to_string(), table: "regional_sales".to_string() }),
    Column::Function("to_date('01/01/2000', 'mm/dd/YYYY')".to_string()),
];
let data: Vec<Vec<SQLDataTypes>> = conn
    .select("regional_sales", columns)
    .inner_join("national_sales", "product_id", "product_id")
    .where_in(product_id, product_ids)
    .and(state, state) 
    .and_not(city, cities)
    .build()?;
data.iter().for_each(|row: &Vec<SQLDataTypes>| { println!("{:?}", row) });

UPDATE

Updates a table's column(s) based on criteria set with an optional conjunction statement. Updates can return Ok() or the number of rows that were updated.

let country = ColumnProps{ name: "country".to_string(), table: "global_sales".to_string() };
let north_american_countries = vec!["Canada".to_sql_fmt(), "United States".to_sql_fmt(), "Mexico".to_sql_fmt()];
let na_countries_formatted = north_american_countries.iter().map(|value| value.to_sql_fmt()).collect::<Vec<SQLDataTypes>>();
let countries = WhereArg::Values(na_countries_formatted);
conn.update("global_sales")
    .set("continent", "North America")
    .where_in(country, countries)
    .build()?;
// If you want to get the number of rows that were updated
let count: usize = conn
    .update("global_sales")
    .set("continent", "North America")
    .where_in(country, countries)
    .build_return_count()?;

INSERT

Inserts a grid (two-dimensional vector) of data into your database. Can take any type that has the ToSQLData trait implemented.

let data: Vec<Vec<&str>> = vec![
    vec!["Column_A", "Column_B", "Column_C"],
    vec!["a1", "b1", "c1"],
    vec!["a2", "b2", "c2"],
    vec!["a3", "b3", "c3"],
];

conn.insert("my_table", data)?.build()?;

If the table does not exist, you can add the create_table() method to automatically create the table.

conn.insert("my_table", data)?.create_table().build()?;

If you have a grid of strings that have integers, dates, etc.. that you want to be formatted properly before being inserted into a table then you want to add the .format_grid_strings() method.

let data: Vec<Vec<&str>> = vec![
    vec!["product_id", "date_sold", "price"],
    vec!["P-001", "09/15/2024", "54.99"],
    vec!["P-002", "10/22/2024", "19.99"],
    vec!["P-003", "11/04/2024", "39.99"],
];

conn.insert("sales_data", data)?
    // Will convert the "date_sold" column into chrono::NaiveDateTime
    // and the "price" column into f64.
    .format_grid_strings()?
    .build()?;

CREATE

Creates a table using a vector of the CreateColumns struct and the CreateDataTypes to apply the correct types to the new columns.

let columns = vec![
        CreateColumns{ name: "Column_A".to_string(), data_type: CreateDataTypes::VARCHAR(20 as usize) },
        CreateColumns{ name: "Column_B".to_string(), data_type: CreateDataTypes::NUMBER },
        CreateColumns{ name: "Column_C".to_string(), data_type: CreateDataTypes::FLOAT },
    ];

    conn.create()
        .table("my_table", columns)
        .build()?;

You can add a column after you initiated the create table process.

let my_table: CreateTable = conn.create()
    .table("my_table", columns);

if add_date_col == true {
    my_table.add_column("Column_D".to_string(), CreateDataTypes::DATE);
}

my_table.build()?;

DELETE

Deletes rows in a table based on the where methods added to the DeleteProps. If no where methods are added, it will delete all data in the table.

let status = ColumnProps{ name: "status".to_string(), table: "employee_data".to_string() };
let terminated = WhereArg::Query("SELECT status FROM employment_statues WHERE status_type LIKE 'termina%'");
conn.delete("employee_data")
    .where_in(status, terminated)
    .build()?;

ALTER

Alters a table by renaming it or adding, modifying, dropping, or renaming a column.

// renaming a table
conn.alter()
    .table("local_sales")
    .rename("regional_sales")
    .build()?;

// Adding a column
let column = AlterColumns {
    name: String::from("title"),
    data_type: CreateDataTypes::VARCHAR(10),
    default: Some(String::from("PMO")),
    not_null: true,
};
conn.alter()
    .table("employees")
    .add(vec![column])
    .build()?;

// Modifying a column (very similar to adding)
conn.alter()
    .table("employees")
    .modify(vec![column])
    .build()?; 

// Dropping a column
conn.alter()
    .table("sales")
    .drop("description")
    .build()?;

// Renaming a column
conn.alter()
    .table("sales")
    .rename_column("salesman", "employee")
    .build()?;

Where

Conjunction statements are split into 4 categories via the WhereArg enum to prevent SQL injections, potential issues with NULL values, and for more intentional query structure.

• Values is a vector of any data type that has is implemented for the ToSQLData trait. This would be used as if you have a basic WHERE clause that you have set values for.

.where_in(column, WhereArg::Values(vec!["one".to_sql_fmt(), "two".to_sql_fmt(), "three".to_sql_fmt()]))

WHERE column IN ('one', 'two', 'three');

• Like is used the same way as a SQL LIKE statement.

.where_in(column, WhereArg::Like("Hello Wor%"))

WHERE column IN ('Hello Wor%');

• Query is when you want to select values from another table.

.where_in(column, WhereArg::Query("SELECT value FROM another_table"))

WHERE column IN (SELECT value FROM another_table)

• NULL is for selecting NULL values.

.where_in(column, WhereArg::NULL)

WHERE column IS NULL

ToSQLData

ToSQLData is the trait that is used to convert various data types to SQLDataTypes.

To implement a local enum:

enum MyEnum {
    Name(String),
    Age(i64)
}

impl ToSQLData for MyEnum {
    fn to_sql_fmt(&self) -> SQLDataTypes {
        match self {
            MyEnum::Name(val) => SQLDataTypes::Varchar(val.into()),
            MyEnum::Age(val) => SQLDataTypes::Number(val.into()),
        }
    }
}

To implement a foreign enum:

use some_crate::SomeForeignType;

struct MyType<'a>(&'a SomeForeignType);

impl ToSQLData for MyType<'_> {
    fn to_sql_fmt(&'_ self) -> SQLDataTypes {
        match self {
            SomeForeignType::Int(val) => SQLDataTypes::Number(*val),
            SomeForeignType::Float(val) => SQLDataTypes::Float(*val),
            SomeForeignType::String(val) => SQLDataTypes::Varchar(val.to_owned()),
            SomeForeignType::None => SQLDataTypes::NULL,
        }
    }
}