Project

General

Profile

« Previous | Next » 

Revision 7395

Added by Aaron Marcuse-Kubitza about 12 years ago

sql_io.py: put_table(): Documented that much of the complexity of the normalizing algorithm is due to PostgreSQL not having a native command for insert/on duplicate select

View differences:

lib/sql_io.py
342 342 
    '''Recovers from errors.
343 343 
    Only works under PostgreSQL (uses INSERT RETURNING).
344 344
345 
    Warning: This function's insert/on duplicate select algorithm does not
346 
    support database triggers that populate fields covered by a unique
347 
    constraint. Such fields must be populated by the mappings instead.
345 
    Warning: This function's normalizing algorithm does not support database
346 
    triggers that populate fields covered by a unique constraint. Such fields
347 
    must be populated by the mappings instead.
348 348 
    (Other fields are not affected by this restriction on triggers.)
349
350 
    Note that much of the complexity of the normalizing algorithm is due to
351 
    PostgreSQL (and other DB systems) not having a native command for
352 
    insert/on duplicate select. This operation is a cross between MySQL's
353 
    INSERT ON DUPLICATE KEY UPDATE (which does not provide SELECT
354 
    functionality), and PostgreSQL's INSERT RETURNING (which throws an error
355 
    instead of returning the existing row).
349 356 
    @param in_tables The main input table to select from, followed by a list of
350 357 
        tables to join with it using the main input table's pkey
351 358 
    @param mapping dict(out_table_col=in_table_col, ...)

Also available in: Unified diff