SQLAlchemy: does not call __init__

June 14, 2009 § Leave a comment

When performing query(), SQLAlchemy does not execute __init__ of the corresponding ORM objects.

Thus, if you have some logic inside __init__, those won’t get executed.

To have the desired behavior, you need to put such logic inside a function that takes no arguments, then, attach @orm.reconstructor on it.



SqlAlchemy Quickie: How to use Distinct

May 15, 2009 § Leave a comment

distinct() method is accessible from Query object. The documentation is kind of obscure, so hopefully this post is helpful for others.


def get_all_category(cls):
return meta.Session.query(cls.category).distinct().all()


SqlAlchemy 04 Documentation about Query object

SqlAlchemy in Three Minutes

May 10, 2009 § Leave a comment

The documentations of SqlAlchemy gets better in every version, but I can never remember the steps of creating Engine, Index, and Session.

So, I created 3 minutes tutorial on how to define the ORM class using declaration_base(), build that table, populates, and fetch.

See code here.

SQLObject: Creating table dynamically

October 25, 2008 § Leave a comment

I’ve been using SQLObject for a while now, and I’m 60/40 with it.

I am mostly happy of SQLObject because:

  • The documentation is easy to read.
  • Getting up to speed is ridiculously easy, in less than 5 minutes I have a ‘user’ MySQL table wrapped with class User.
  • The generated SQL is efficient, for all my needs.
  • The returned ResultSet can be sliced, that’s nice. 

Why I’m not happy with it:

  • The invisible ‘id‘ column, that always get created for every new table. I have multiple needs for ORM to just create the table as-is, explicitly (Like how SQLALchemy does).

Why I haven’t switched to SQLAlchemy:

  • I have, but not for all model classes.
  • My noobiness limits me on what I can do with SQLAlchemy. For example, I haven’t figure out how to add Column object to existing Table object (that was created dynamically).  

That being said, SQLObject is very useful if you want to create MySQL table dynamically. You can achieve this by using meta-programming.


def create_or_get_dynamic_mysql_table(name):
      # SQLObject cache all classes that already defined,
      # If 'name' already defined, just get it using findClass()
      dynamic_table_classname = findClass(name)
      # Create new 'name' class
      # If you see the 2nd param of classobj, 'name' class extends SQLObject class
      dynamic_table_classname = classobj( name, (sqlobject.SQLObject,), {} )

   return dynamic_table_classname

# Then Call the createTable method
table_class = create_or_get_dynamic_mysql_table('awesome')


SqlAlchemy: SQLError: (OperationalError) (2006, ‘MySQL server has gone away’)

March 2, 2008 § Leave a comment

What is this problem about:

This is actually MySQL error message. It means that there are no MySQL connections anymore.By MySQL default, idling database connections will be closed after 8 hours. Not having MySQL connections means that no SQL operations can be performed. SqlAlchemy is simple reporting the error.

When using SQLAlchemy, how to prevent such problems?

See sample code below:

engine = create_engine(db_path, pool_size = 100, pool_recycle=7200)

 Take a look at the pool_recycle variable. That will configure SqlAlchemy engine to reconnect every x number of seconds. This easy solution makes sure the existence of MySQL connection.


Why Python?

January 20, 2008 § 3 Comments

Among other programming languages, why choose Python?

Let’s go over the basics first, these traits also exist in other programming languages:

  • Python is interpreted Programming Language. It is also a high-level programming language. Therefore, just like Ruby or Perl, Python is easy to learn and easy to get up to speed on sizable projects.
  • Python is also object-oriented language. But that’s nothing new, most modern programming languages are object-oriented anyway.

Naturally, readers will ask, if Python can only do what MY programming language already can… Why do I have to learn it? Below is my reasoning why… and hopefully readers will convert to Python 😛

1. Python does not have funky sigils or special system_call functions. The only one you have to memorize is: __someDefaultMethods__ By not having too many special characters, programmers spend less time browsing documentation or mailing list on Google.

2. Python tabbed indentation actually contributes to highly readable code. Readable code helps especially when you are juggling many different projects every week, it allows you to remember quickly what you did last time.

3. Python’s power and scalability is proven not hype.

4. For Back-end Programmers, Database Designer, and System Engineers:

  • Look at Twisted[1]. It is one kick ass networking library. It supports many networking protocols that allows building web servers, email, or chat easier.
  • Look at SQLAlchemy[2][3], I never seen ORM so powerful and yet so easy to learn.

5. For Application Developers:

  • Django is a complete out-of-the-box Web Framework. Their slogan — “For Perfectionists with deadlines”
  • TurboGear is also another awesome complete Web Framework solution. But I suggest readers wait until they included SQLAlchemy as their new database stack.

6. For Javascript Developers:

  • Look at Mochikit, It is a complete javascript framework that’s very Pythonic. Mochikit is still just Javascript, but have Python’s sense of elegance.

Below are quotes from Eric Raymond on why he likes Python:

— “In Python, I was actually dealing with an exceptionally good design. Most languages have so much friction and awkwardness built into their design that you learn most of their feature set long before your misstep rate drops anywhere near zero. Python was the first general-purpose language I’d ever used that reversed this process.

— “It (Python) is compact–you can hold its entire feature set (and at least a concept index of its libraries) in your head. C is a famously compact language. Perl is notoriously not;

— “What I really wanted was code that would analyze the shape and members of the initializer, query the class definitions themselves about their members, and then adjust itself to impedance-match the two sets… …This kind of thing is called metaclass hacking… …Thirty-two lines, counting comments… …Brace yourself: this code only took me about ninety minutes to write—and it worked correctly the first time I ran it.

— “So the real punchline of the story is this: weeks and months after writing fetchmailconf, I could still read the fetchmailconf code…


Hopefully I provided enough examples on why Python can be your next favorite programming language. Programming in Python is fun indeed.


  1. http://twistedmatrix.com/trac/wiki/FrequentlyAskedQuestions
  2. http://www.sqlalchemy.org/
  3. http://en.wikipedia.org/wiki/SQLAlchemy
  4. If you are new in Programming, this E-book can help
  5. Why use Python in Windows XP?
  6. Python Doc – String
  7. Python Doc – Common OS path manipulation
  8. If you are familiar with apt-get, Python have something similar for its modules: easy_install

What Others Said about Why Python:

Interesting criticism against SQLAlchemy

December 30, 2007 § Leave a comment

The Hand of FuManchu pits SQLAlchemy against Storm and Geniusql under his benchmarks.

In the INSERT benchmark FuManchu(I believe he is involved with CherryPy) claims that SQLAlchemy is making over 400 function calls.

FuManchu show these 2 findings on his benchmarks:

  • SQLAlchemy spends a lot of time building the bound parameters.
  • He also claims that SQLAlchemy takes 73 Python function calls fetching auto-increment id.

In response, Mike Bayer wrote, in his blog, saying that SQLAlchemy aims to be a complete ORM with rich features. He even corrected the benchmarks here.

What does this benchmarks mean to me? Nothing much really. I like using SQLAlchemy, it has a lot of features I need and some more. IF this benchmark is right, I can always use raw SQL just for the sake of speed, although such code is ugly.

For readers, read the blogs comments as well. They are very interesting.

Where Am I?

You are currently browsing the sqlalchemy category at RAPD.