You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Current implementation uses QuerySet slicing, which translates into a LIMIT n OFFSET x in SQL.
The problem is that your database has to go through your entire table to select only a small quantity of rows. If your database contains thousands or millions of rows, this leads to far slower queries.
That would mean switching from qs[offset:limit] to qs.filter(**kwargs)[:limit] where kwargs is a dict created from the ORDER BY columns and their values in the previous page.
A Django implementation of this solution already exists. But it’s not using the ORM to discover which columns are ordering, and you can only order by one column, so it’s quite useless.
The columns in ORDER BY can be easily fetched using this:
q=qs.queryorder_by=q.extra_order_byor (q.get_meta().orderingifq.default_orderingelseq.order_by) or []
Of course, the real challenge is to get data from the previous page. Unless we already asked for the previous page, we can’t get these data. In that "first query scenario", nothing special can be done, we have to use OFFSET. Otherwise, there are two solutions in my opinion:
pass these data in the template, so that the GET request of the next page contains data from the previous page (easy but less efficient)
save these data to the cache for each page, so that it’s saved from a session to another; this implies having to invalidate this cache after each modification (harder but more efficient)
The text was updated successfully, but these errors were encountered:
Current implementation uses QuerySet slicing, which translates into a
LIMIT n OFFSET x
in SQL.The problem is that your database has to go through your entire table to select only a small quantity of rows. If your database contains thousands or millions of rows, this leads to far slower queries.
This can be solved using keyset pagination.
That would mean switching from
qs[offset:limit]
toqs.filter(**kwargs)[:limit]
wherekwargs
is a dict created from theORDER BY
columns and their values in the previous page.A Django implementation of this solution already exists. But it’s not using the ORM to discover which columns are ordering, and you can only order by one column, so it’s quite useless.
The columns in
ORDER BY
can be easily fetched using this:Of course, the real challenge is to get data from the previous page. Unless we already asked for the previous page, we can’t get these data. In that "first query scenario", nothing special can be done, we have to use
OFFSET
. Otherwise, there are two solutions in my opinion:The text was updated successfully, but these errors were encountered: