Details
-
Task
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Won't Fix
-
None
Description
Hello,
I have a customer who has run into an issue when they have data they want to query in information_schema tables but the tables have 100,000's+ records in them. The more records that get added the worse their query performs.
Could it be possible to set up a way to allow indexes be created against tables in the information_schema? Whether it be allowing
alter table add index... |
or some way to specify it in the MariaDB config file.
These indexes should not be created by default, and only by users who feel they need them.
Attachments
Issue Links
- relates to
-
MDEV-11655 Transactional data dictionary
-
- Open
-
-
MDEV-17581 Data dictionary
-
- Open
-
-
MDEV-20609 Full table scan in INFORMATION_SCHEMA.PARAMETERS/ROUTINES
-
- Closed
-
Because of the way the information_schema is written, its currently easier to write in specific optimizations like those that exists for information_schema.tables when in a specific database/table.
If you provide a more specific information_schema table request and what sort of look ups are being performed maybe an optimization can be achieved in a shorter development time.
Assuming this request is table related, the alternative is for the data dictionary aspects to be implemented and for this to become a real table, which would have greater support for arbitrary indexes.