Try this: http://stackoverflow.com/questions/18412540/sqlalchemy-order-by-formula-result

Hi,

There are two options for this: in the query, or in memory with pandas.

The DB version of this is:

fundamental_df2 = get_fundamentals(  
        query(  
            fundamentals.balance_sheet.total_assets,  
            fundamentals.balance_sheet.current_assets,  
            fundamentals.balance_sheet.goodwill_and_other_intangible_assets,  
            (fundamentals.balance_sheet.total_assets -\  
             fundamentals.balance_sheet.current_assets -\  
             fundamentals.balance_sheet.goodwill_and_other_intangible_assets).label('nfa') # label is similar to sql's 'select foo as bar'  
            ).order_by('nfa').limit(25)  
    )  

and the in-memory version of the same is:

fundamental_df = get_fundamentals(  
        query(  
            fundamentals.balance_sheet.total_assets,  
            fundamentals.balance_sheet.current_assets,  
            fundamentals.balance_sheet.goodwill_and_other_intangible_assets  
        )  
    ).T # taking the transpose of the normal dataframe, to make symbols into rows for sorting.  
    fundamental_df['nfa'] = fundamental_df['total_assets'] - \  
        fundamental_df['current_assets'] - \  
        fundamental_df['goodwill_and_other_intangible_assets']  
    context.fundamental_df = fundamental_df.sort('nfa')[:200]  
    update_universe(context.fundamental_df.index.values)  

For a simple calculation like this, it is best to push the work to the database in the sql query. But, I wanted to let you know that you can also take uncapped results into memory. That method has the most flexibility, and you can use any of the stats packages etc on the resulting data.

have fun,
fawce

to impose a descending or ascending order on the label we set
).order_by(desc('nfa desc')).limit(25)