问题
If I do something like
SELECT * FROM mytable ORDER BY mycolumn ASC;
I get a result table in a specific order.
Is there a way in SQL to efficiently find out, given a PK, what position in that result table would contain the record with my PK?
回答1:
You can count the number of records where the value that you are sorting on has a lower value than the record that you know the key value of:
select count(*)
from mytable
where mycolumn < (select mycolumn from mytable where key = 42)
回答2:
On databases that support it, you could use ROW_NUMBER() for this purpose:
SELECT RowNr
FROM (
SELECT
ROW_NUMBER() OVER (ORDER BY mycolumn) AS RowNr,
mycolumn
FROM mytable
) sub
WHERE sub.mycolumn = 42
The example assumes you're looking for primary key 42 :)
The subquery is necessary because something like:
SELECT
ROW_NUMBER() OVER (ORDER BY mycolumn) AS RowNr
FROM mytable
WHERE sub.mycolumn = 42
Will always return 1; ROW_NUMBER() works after the WHERE, so to speak.
回答3:
SQL doesn't work that way. It's set-based, which means that "position in that result table" is meaningless to the database.
You can keep track of position when you map the ResultSet into a collection of objects or when you iterate over it.
回答4:
Unfortunately you cannot get "the position of a row in a table".
The best you can get, using ORDER BY and a variant of the ROW_NUMBER construct (depends on the database engine in use), is the position of a row in the resultset of the query executed.
This position does not map back to any position in the table, though, unless the ORDER BY is on a set of clustered index columns, but even then that position might be invalidated the next second.
What I would like to know is what you intended to use this "position" for.
回答5:
There's no way you can tell that without selecting an entire subset of records. If your PK is of integer type, you can
select count(*) from mytable
where id <= 10 -- Record with ID 10
order by mycolumn asc
来源:https://stackoverflow.com/questions/907438/can-i-get-the-position-of-a-record-in-a-sql-result-table