I have decided to rework my local version of the PEAR SQL parser to overcome some limitations.  The main features that I will be adding include:

• Proper handling of sub-selects
• Maintaining order of fields when some fields are actually function calls

In order to do this I have had to add a ‘columns’ array and a ‘tables’ array that hold information about all columns and tables respectively whether they are simple identifiers or not.  That means that some information is duplicated in these arrays (from the column_names and column_aliases, etc.. arrays). 

When this conversion is complete, it will pave the way for surrogate views in Dataface.