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
Currently we use the min/max of pages to determine which pages we need to search. Then we load those pages and search them with a flat search. Previously this was dominated by I/O time. Now that we are caching btree pages this flat search time could be a considerably part of the query.
We can speed this up by adding a fast path for pages that are obviously completely satisfied. For example, if the page has min=7 and max=11 and the query is foo BETWEEN 0 AND 20 then we know the page is completely matched and we can skip the flat search and just add all the row ids.
The text was updated successfully, but these errors were encountered:
Currently we use the min/max of pages to determine which pages we need to search. Then we load those pages and search them with a flat search. Previously this was dominated by I/O time. Now that we are caching btree pages this flat search time could be a considerably part of the query.
We can speed this up by adding a fast path for pages that are obviously completely satisfied. For example, if the page has min=7 and max=11 and the query is
foo BETWEEN 0 AND 20
then we know the page is completely matched and we can skip the flat search and just add all the row ids.The text was updated successfully, but these errors were encountered: