Question

I have been wondering whether there is any code out there that enables representing SQL in the form of some object tree that can be assembled, modified and then finally rendered to valid SQL?

Off the top of my head it could look something like that…

var stmnt = new Statement(); stmnt   .AddMaster('Customer')   .Show('Firstname, 'Lastname')   .AddJoin('Address', 'ID', 'CustomerID')   .Show('Street', 'City'); stmnt.WhereStatement()   .AddParameter('Address.City', Op.Equal);  string sql = stmnt.Generate(); // select a.FirstName, a.LastName, b.Street, b.City // from Customer a // join Address b on b.CustomerID = a.ID // where b.City = :p1 

This is just an example and the thing out there may work totally different, but yes, I’d love to hear what is out tere in that respect.

UPDATE:

I am aware of the numerous possibilities of using ORM technologies to get my results from the DB, but I was after a model for the SQL itself. I know that the level of abstraction is pretty low, but it could possibly allow a situation where multiple collaborators can work on an SQL statement (multiple joins, multiple wheres) which can then be ‘rendered’ at the end of the build-Phase.

Answer 1

The python package SQLAlchemy has an ORM layer, but it also has an SQL generation layer.

[I realise you tagged this post c# and .net, but I thought you might like to see what else is out there]

Here is some example code:

from sqlalchemy import Table, Column, Integer, String, MetaData, ForeignKey from sqlalchemy.sql import select  metadata = MetaData()  # Make a basic customer table. Customer = Table('Customer',                  metadata,                  Column('ID', Integer, primary_key=True),                  Column('FirstName', String),                  Column('LastName', String))  # Make a basic address table Address = Table('Address',                 metadata,                 Column('ID', Integer, primary_key=True),                 Column('City', String),                 Column('Street', String),                 Column('CustomerID', None, ForeignKey('Customer.ID')))   # Generate some sql stmt = select([Customer.c.FirstName,                Customer.c.LastName,                Address.c.Street,                Address.c.City],               from_obj=Customer.join(Address),               whereclause=Address.c.City == 'Wellington')  # Display print stmt # output: SELECT 'Customer'.'FirstName', 'Customer'.'LastName', 'Address'.'Street', 'Address'.'City'  FROM 'Customer' JOIN 'Address' ON 'Customer'.'ID' = 'Address'.'CustomerID'  WHERE 'Address'.'City' = :City_1  # note that SQLAlchemy picked up the join condition from the foreign key. # you can specify other join conditions if you want. 

Typically, you would execute the statement by using SQLAlchemy to connect to a database. Then you can do:

for row in stmt.execute():     print 'Name:', row.c.FirstName, row.c.LastName, 'City:', row.c.City 

Hope this helps.