Reputation: 1932
When we should define db_index=True
on a model fields ?
I'm trying to optimize the application & I want to learn more about db_index
, in which conditions we should use it ?
The documentation says that using db_index=True
on model fields is used to speed up the lookups with slightly disadvantages with storage and memory.
Should we use db_index=True
only on those fields that have unique values like the primary field id
?
What happens if we enabled indexing for those fields which are not unique and contains repetitive data ?
Upvotes: 35
Views: 28329
Reputation: 1778
When you set db_index=True
on some field, queries based on that field would be much faster O(log(n)) instead O(n).
Under the hood, it is usually implemented using B-Tree.
The trade-off for these accelerated queries is increased memory usage and time for writes. So the best use case for indexing would be if you have a read-heavy table that is often queried by non-primary field.
Upvotes: 13
Reputation: 458
I would say you should it when you have a field that is unique for fast lookups.
e.g., You have a table of customers
with manyusers
. Each row representing a user will have their own unique user_id
. When you create an index, a pointer is created within the DB to where that data is stored so that the next look up won't take as long.
Have a look here to learn more
Upvotes: 28
Reputation: 983
You should use db_index=True
when you use unique=True
, there is a specific reason to use it,
By using this method you can boost a little bit of performance,
When we fire a query in SQL, finding starts from the top to bottom
Case: 'Without db_index=True': It will search and filter till all bottom rows even if we find the data
Case: 'With db_index=True': When Object finds it will just stop their
It will boost a little bit of performance
Upvotes: 7