Project

General

Profile

« Previous | Next » 

Revision 3099

sql.py: distinct_table(): Return new table instead of renaming input table so that columns that use input table will continue to work correctly

View differences:

lib/sql.py
1226 1226

  
1227 1227
def distinct_table(db, table, distinct_on):
1228 1228
    '''Creates a copy of a temp table which is distinct on the given columns.
1229
    The table will get an index on these columns, facilitating merge joins.
1230
    @param table Will be renamed to the distinct on table.
1229
    The old and new tables will both get an index on these columns, to
1230
    facilitate merge joins.
1231 1231
    @param distinct_on If empty, creates a table with one row. This is useful if
1232 1232
        your distinct_on columns are all literal values.
1233
    @return The old table.
1233
    @return The new table.
1234 1234
    '''
1235
    old_table = copy.copy(table)
1236
    table.name = sql_gen.concat(table.name, '_distinct')
1235
    new_table = sql_gen.suffixed_table(table, '_distinct')
1237 1236
    
1238
    copy_table_struct(db, old_table, table)
1237
    copy_table_struct(db, table, new_table)
1239 1238
    
1240 1239
    limit = None
1241 1240
    if distinct_on == []: limit = 1 # one sample row
1242
    else: add_index(db, distinct_on, table, unique=True)
1241
    else:
1242
        add_index(db, distinct_on, new_table, unique=True)
1243
        add_index(db, distinct_on, table) # for join optimization
1243 1244
    
1244
    insert_select(db, table, None, mk_select(db, old_table, start=0,
1245
    insert_select(db, new_table, None, mk_select(db, table, start=0,
1245 1246
        limit=limit), ignore=True)
1246
    analyze(db, table)
1247
    analyze(db, new_table)
1247 1248
    
1248
    return old_table
1249
    return new_table

Also available in: Unified diff