List:Summer of Code« Previous MessageNext Message »
From:scut_tang Date:July 29 2009 11:29am
Subject:Some questions about get_all_tables
View as plain text  
Hi, Sergei
  I have understanded how get_all_tables() work roughly. The index for tables based on
filesystem is to be implemented by making the right database names and table names. And
the process_table function pointer is to fill data into table according by database name
and table name.
The old_format function pointer is used for SHOW statements, also there are some old names
in fileds_info. I want do know more about MySQL internal, instead of being a code porter
:) The questions list:
1.  What does thd->lex->wild mean? I see there are lots of "char *wild" and
functions, like wild_case_compare(). Especially, the struct 
LOOKUP_FIELD_VALUES, which contains wild_db_name and wild_table_value;
2. What do table->next_number_field and table->found_next_field mean?
3. The I_S storage engine should support SHOW statements? like, SHOW COLUMNS FROM XXX.XXX;
4. In get_all_tables(), SELECT_LEX *lsel= tables->schema_select_lex comes from
TABLE_LIST *tables, but there are no TABLE_LIST in class handler.
But I see table->pos_in_table_list and table->pos_in_locked_tables. Could I use
them?
5. Why storage engine do not get some fields' data according by table->read_set,
instead of all fields. For example:
  SELECT id FROM t1; only get id field data, not entire record. 
6. What is the purpose of partial_cond->val_int() in get_all_tables()? Verify the
condition again?
 
So many questions :P
Regards,
Robin


Thread
Some questions about get_all_tablesscut_tang29 Jul