sqlgen/docs/to_sql.md

sqlgen::...::to_sql

The to_sql functions in each dialect namespace (e.g., sqlgen::postgres::to_sql, sqlgen::sqlite::to_sql) provide a way to transpile C++ SQL operations into dialect-specific SQL strings. This is particularly useful for testing, debugging, logging, and understanding the exact SQL that will be executed by sqlgen.

Usage

Basic Usage

Transpile any SQL operation to a dialect-specific SQL string:

using namespace sqlgen;
using namespace sqlgen::literals;

// Define a query
const auto query = sqlgen::read<std::vector<Person>> |
                   where("age"_c < 18 and "first_name"_c != "Hugo") |
                   order_by("age"_c);

// Transpile to PostgreSQL dialect
const auto postgres_sql = postgres::to_sql(query);

This generates the following PostgreSQL SQL:

SELECT "id", "first_name", "last_name", "age" 
FROM "Person" 
WHERE ("age" < 18) AND ("first_name" != 'Hugo') 
ORDER BY "age";

Or for SQLite:

// Transpile to SQLite dialect
const auto sqlite_sql = sqlite::to_sql(query);

Supported Operations

to_sql supports transpilation of all sqlgen operations:

SELECT Queries

using namespace sqlgen;
using namespace sqlgen::literals;

// Basic select
const auto select_query = sqlgen::read<std::vector<Person>>;
const auto sql = postgres::to_sql(select_query);

This generates the following SQL:

SELECT "id", "first_name", "last_name", "age" 
FROM "Person";

With a where clause:

// With where clause
const auto filtered_query = sqlgen::read<std::vector<Person>> | 
                           where("age"_c >= 18);
const auto sql = postgres::to_sql(filtered_query);

This generates the following SQL:

SELECT "id", "first_name", "last_name", "age" 
FROM "Person" 
WHERE "age" >= 18;

CREATE TABLE Operations

Note that every call to sqlgen::write calls a CREATE TABLE operation before actually inserting the data.

using namespace sqlgen;
using namespace sqlgen::literals;

// Define a table structure
struct TestTable {
    std::string field1;
    int32_t field2;
    sqlgen::PrimaryKey<uint32_t> id;
    std::optional<std::string> nullable;
};

// Create table query
const auto create_query = create_table<TestTable> | if_not_exists;
const auto sql = postgres::to_sql(create_query);

This generates the following PostgreSQL SQL:

CREATE TABLE IF NOT EXISTS "TestTable" (
    "field1" TEXT NOT NULL,
    "field2" INTEGER NOT NULL,
    "id" INTEGER NOT NULL,
    "nullable" TEXT,
    PRIMARY KEY ("id")
);

INSERT Operations

using namespace sqlgen;
using namespace sqlgen::literals;

const auto insert_query = Insert<TestTable>{};
const auto sql = postgres::to_sql(insert_query);

This generates the following PostgreSQL SQL:

COPY "public"."TestTable"("field1", "field2", "id", "nullable") 
FROM STDIN WITH DELIMITER '\t' NULL '\e' QUOTE '\a';

For SQLite, the same operation generates:

INSERT INTO "TestTable"("field1", "field2", "id", "nullable") 
VALUES (?, ?, ?, ?);

DELETE Operations

using namespace sqlgen;
using namespace sqlgen::literals;

const auto delete_query = delete_from<TestTable> | 
                         where("field2"_c > 0);
const auto sql = postgres::to_sql(delete_query);

This generates the following SQL:

DELETE FROM "TestTable" 
WHERE "field2" > 0;

Dialect-Specific Features

Each database dialect implements its own version of to_sql with dialect-specific features:

PostgreSQL (sqlgen::postgres::to_sql)

• Uses COPY for bulk inserts
• Supports schema names (e.g., "public")
• Uses PostgreSQL-specific escaping rules

SQLite (sqlgen::sqlite::to_sql)

• Uses INSERT INTO for inserts
• Uses SQLite-specific escaping rules
• Simpler schema (no schema names)

Notes

• Each dialect's to_sql is a template function that works with any sqlgen operation
• The functions are noexcept and return a std::string
• The generated SQL is properly escaped and safe to use
• The functions are particularly useful for:
  • Debugging complex queries
  • Logging SQL operations
  • Understanding the exact SQL that will be executed
  • Testing and validation
• Each dialect's implementation ensures the generated SQL is compatible with that specific database
• The functions preserve all SQL operations' semantics while adapting to dialect-specific syntax