Hello,
I am trying to understand how the cost for a query involving indexes on expressions is calculated. How is the statistics on the _expression_ maintained? For example Postgres documentation on 'Indexes on Expressions' mentions the following example:
CREATE INDEX people_names ON people ((first_name || ' ' || last_name));
SELECT * FROM people WHERE (first_name || ' ' || last_name) = 'John Smith';
The index is created on an _expression_ involving first name and last name. My confusion is basically around:
(1) When using this index how will the query planner get the stats on this _expression_?
(2) Whenever an index on an _expression_ is defined like this, should we rebuild the stats?
Thanks,
Shantanu
