sql.py: pkey_col(): Specify recover directly as a kw_arg because it's the only kw_arg passed to pkey_name()

sql.py: Added table_pkey_col() and use it in pkey_name()

sql.py: Renamed pkey() to pkey_name()

sql.py: Renamed pkey_col_() to pkey_col()

sql.py: Removed no longer used pkey_col

db_xml.py: cleanup_table(): Inline sql.pkey_col ('row_num') because this is the only place it's used

cleanup_table(): Use new sql.table_cols() instead of sql.table_col_names()

sql.py: Added table_cols()

db_xml.py: put(): Fixed bug where needed to avoid truncating the pkeys_loc table, in case it's the same as one of the in_tables. This occurs now that sql_io.put_table() passes through the actual input column instead of the joined-together input table's column when ignoring all rows.

sql_io.py: put_table(): Resolving default value column: If ignoring all rows, use input cols directly instead of cols from joined-together input table. In addition to being simpler, this prevents the returned column's name from growing longer and longer as each iteration prepends its input table table name to the default value column name.