Tracker

HOME

TRACKER

MYSQL

Subselect

Unless I'm wrong, here's a way to do a subquery (inner join two tables, then inner join the resulting table with a third table). It takes advantage of the two different ways of expressing an inner join ("INNER JOIN", and "t1, t2 WHERE...") to express two separate inner joins within a single statement.
SELECT p.p_id, v2.v_name FROM t_project p, t_volunteer v INNER JOIN t_volunteer v2 ON p.p_id=v2.p_id WHERE p.p_id=v.p_id AND v.v_name LIKE "%mike%";
Is this a technique that people use often? I couldn't see it documented in my SQL book ("MySQL", by Paul DuBois), even though it seems like a useful technique for what is effectively a subselect.

View Complete Forum Thread with Replies

See Related Forum Messages: Follow the Links Below to View Complete Thread

Getting Rid Of Subselect
I have a table, which -- simplified -- looks like this:

create table access_logs (
session_id varchar(32),
request_uri varchar(32)
);
Each pageview logs the users session-id + the request-uri. Now, to determine how many visitors followed a specific path, I need to select the number of sessions, which have a row including specific request_uri.

This is my own feeble attempt, but I have a feeling that this could be rewritten to get rid of the subselects:

select session_id
from access_logs
where session_id in (select session_id
from access_logs
where request_uri = "landing-page.php")
and session_id in (select session_id
from access_logs
where request_uri = "exit-page.php")
group by session_id;

Getting Around Subselect
My knowledge of SQL is basic so I need some help developing a query. Suppose we have a table called BID where each row is a bid on an item up for auction. The relevant columns are BID_ID which is the primary key, ITEM_ID which identifies the item, and a BID_DATE which records the datetime of the bid.

I would like to find the most recent bid for each distinct ITEM_ID in the table. I've worked out the query below which seems to do the job. However, I need to find a query that will work on a pre-4.1 server which does not support subqueries. Is there a way to re-state this query without using a subselect? Perhaps using some kind of join?

SELECT *
FROM BID, (SELECT ITEM_ID AS IID, max(BID_DATE) as MAXDATE
FROM BID
GROUP BY ITEM_ID) as MAXDATES
WHERE
(ITEM_ID=IID) and (BID_DATE=MAXDATE);

Subselect
I had some SQL calls which worked fine on a v4.1 server and now I've moved to another one which is 4.0.24 and certain subselects no longer work. Is there any basic way to convert statements such as this:

SELECT a.name, (SELECT COUNT(*) FROM table2 AS b WHERE b.id=a.id) as count FROM table1 AS a So that it conforms to the 4.0.x standard?

Subselect
i am trying to remove values from a list menu if the join table
doesnt have keys when a key is selected for instance:

locations
locationID

locations_join
locationID shotlistID

SELECT SQL_NO_CACHE l.locationID , l.location FROM locations l LEFT JOIN
locations_join lj ON l.locationID = lj.locationID WHERE l.locationID NOT IN
(select locationID FROM locations_join WHERE shotlistID IN (5069)) ORDER BY
l.location ASC

so when shotlistID is selected all the keys from the locations_join joined
to the shotlistID would be remove from the locations list please help, i'm
trying to do this in one query saving from getting all the keys into an
array then checking if the values arent in the array when generating the
list. Code:

Subselect
Let's say i got this query:
select user.id, (select count(*) as posts_number from posts where posts.user_id = user.id), and some other fields, and a lot of joins here.Is it any way to *say* to mysql that the current user.id selected, is the one in the subselect ?
(the one from select "user.id" and the one in where ... = "user.id")

Subselect
I have been held up long enough on the query time to ask for help. Its basically a subselect that never returns.

SELECT id, it.org_id FROM import_temp3 AS it WHERE it.org_id IN ( SELECT p.org_id FROM join_to_person AS j, person AS p WHERE p.id = j.person_id AND j.value = '15' ORDER BY p.org_id ASC ) ORDER BY it.org_id ASC

If I break it up into 2 seperate

SELECT id, it.org_id FROM import_temp3 AS it WHERE it.org_id = 09238323 ORDER BY it.org_id ASC

SELECT p.org_id FROM join_to_person AS j, person AS p WHERE p.id = j.person_id AND j.value = '15' ORDER BY p.org_id ASC

They both return expected values.

Subselect / AS
SELECT
a.id
(SELECT width, height, filename FROM photos WHERE user_id = a.id LIMIT 0,1) AS (width, height, filename)
FROM users a

ORDER BY a.datestamp DESC

you can see my example, using with AS (example).
How can i extract values from subselects?

Slow Subselect
I've got two tables:
lo_users: nickname|id|...
lo_friends: from|to|...

The following query takes < 0.01 sec:
SELECT IF(`from` = '10855', `to`, `from`) userid FROM lo_friends WHERE (`from` = '10855' OR `to` = '10855') AND STATUS = '1'

...but if I use it in a subselect, the whole thing takes about 0.54 sec:
SELECT u.nickname FROM (SELECT IF(`from` = '10855', `to`, `from`) userid FROM lo_friends WHERE (`from` = '10855' OR `to` = '10855') AND STATUS = '1') f LEFT JOIN lo_users u ON u.id = f.userid

What can I do to make the query faster? "from" and "to" are indexed and lo_users.id is the primary key.

Subselect In 4.0.12-max With -- New Option
I was reading the manual and it said that the subselect is only
available in 4.1 or using the 4.0.12 with the mysqld =96new command line
to start it.

But it doesn=92t working!! So I downloaded the 4.1 alpha version with =
the
same problem. The error is:

ERROR 1064: You have an error in your SQL syntax. Check the manual that
corresponds to your MySQL server version for the right syntax to use
near =85.

Any ideas? I need subselect working.

Delete Subselect
I know that MySQL 3.23.nnn did not support a delete subselect, just wondering
what the best/most efficient way to do the following is:

delete from table_a where table_a.column_1 in ( select column_1 from table_b);

Assuming that column_1 is the same data type and size in both table_a and table_b.

Update Self Subselect
I have a log table that creates a row for every page view. I have a field called "flagged" which defaults to 0. I'm trying to update the flagged field to 1 when the ip count is greater then 30... This is what I have but I get the error "You can't specify target table 'ip_log' for update in FROM clause".

UPDATE `ip_log` SET `flagged` = 1
WHERE `ip` IN(
SELECT `ip`
FROM `ip_log`
GROUP BY `ip` HAVING COUNT(*) > 30
)

Subselect Wierdness
I am trying to get 3 active article IDs from the table ARTICLES for a random active feed from table FEEDS. Here is the query I have:

CODEselect AID, a.fid as FID from ARTICLES a where active='Y' and a.fid = (SELECT f.fid FROM FEEDS f where active='Y' ORDER BY RAND() desc limit 1) limit 3;

Subselect With NULL
why don't i get some results for the second query?

mysql> select * from a;
+---+
| b |
+---+
| a |
| b |
+---+

mysql> select * from a where b not in (select NULL from dual);
Empty set (0.00 sec).

Subselect Doesnt Work
i am trying to remove values from a list menu if the join table
doesnt have keys when a key is selected for instance:
locations
locationID

locations_join
locationID shotlistID

SELECT SQL_NO_CACHE l.locationID , l.location FROM locations l LEFT JOIN
locations_join lj ON l.locationID = lj.locationID WHERE l.locationID NOT IN
(select locationID FROM locations_join WHERE shotlistID IN (5069)) ORDER BY
l.location ASC

so when shotlistID is selected all the keys from the locations_join joined
to the shotlistID would be remove from the locations list please help, i'm
trying to do this in one query saving from getting all the keys into an
array then checking if the values arent in the array when generating the
list.

Get Last Records Details With A Subselect?
I have a ticketsystem where each ticket belongs to an user and each user can insert a couple of messages to one ticket. Therefore I have implemented a date field (used as primary key). Now I want to get details from the last entry belongs to a ticketid.

kdn_message:
updated (date)
ticketid (int)
kdnr (int)
detail
state

select * from kdn_message t
where updated in (select max(updated) as updated from kdn_message group by ticketid where ticketid=t.ticketid order by updated desc)

what is wrong i this statement?

Convert A Subselect To Inner Join
I was developing a php/postuke app for a client and I wrote two of my SQL queries with subselects. I found out after I was done that they were pretty much stuck with MySQL 4.0.x for awhile, so I need to revamp my queries to avoid subselects. The query uses three tables:

nuke_gwbt_guild_halls
nuke_gwbt_guild_halls_notes
nuke_gwbt_matches

I am getting all of the fields in the first table, matching the notes id from the second table to a notes id in the first table, and then counting some metrics from the third table to return as fields in the resulting recordset, used for ORDER BY sorting. Here is the working subselect query: Code:

Subselect Doesnt Work
i am trying to remove values from a list menu if the join table d=
oesnt have keys when a key is selected for instance:
locations
locationID

locations_join
locationID shotlistID

SELECT SQL_NO_CACHE l.locationID , l.location FROM locations l LEFT JOIN lo=
cations_join lj ON l.locationID =3D lj.locationID WHERE l.locationID NOT IN=
(select locationID FROM locations_join WHERE shotlistID IN (5069)) ORDER B=
Y l.location ASC

so when shotlistID is selected all the keys from the locations_join joined =
to the shotlistID would be remove from the locations list please help, i'm =
trying to do this in one query saving from getting all the keys into an arr=
ay then checking if the values arent in the array when generating the list.

Subselect/left Join
I have a table like this

| ID | THING | NUMBER |
---------------------------------------------------------------
| 1 | white | 1 |
| 2 | white | 2 |
| 3 | green | 1 |
| 4 | green | 3 |
| 5 | brown | 1 |
| 6 | brown | 4 |

and I want to get just white back if I know two numbers are 1 and 2 or green back if I know the nubmers are 1 and 3.

Its mysql 4.1 so I am allowed subselects or left joins. I am drawing a blank!?

Complex Select (Possible Subselect Needed?)
I have a table, b5_assignment_lookup, that is used elsewhere as a lookup but I'm trying to use the data contained by itself here. The table:

CREATE TABLE b5_assignment_lookup (
as_id int(11) NOT NULL auto_increment,
as_blog int(11) NOT NULL default Ɔ',
as_blogger int(11) NOT NULL default Ɔ',
as_milestone enum('start','finish') NOT NULL default 'start',
as_timestamp timestamp NOT NULL default CURRENT_TIMESTAMP,
PRIMARY KEY (`as_id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=222 ;
By running this query, I get the following resultset:

mysql> SELECT * FROM b5_assignment_lookup WHERE as_blog = ྕ' AND as_timestamp <= &#55614;&#57158;-09-30'
+--------+----------+-------------+---------------+---------------------+
| as_id | as_blog | as_blogger | as_milestone | as_timestamp |
+--------+----------+-------------+---------------+---------------------+
| 87 | 89 | 41 | start | 2006-05-01 00:00:00 |
|208 | 89 | 41 | finish| 2006-09-02 11:55:27 |
|209 | 89 | 103 | start | 2006-09-02 11:55:27 |
+--------+----------+-------------+---------------+---------------------+
3 rows in set (0.00 sec)
What we have here is that Blogger 41 began writing on Blog 89 on May 1, and on Sep 2 Blogger 103 took over for Blogger 41.

What I really need to grok is all bloggers who blogged all or a part of a range of dates.

For example, I really want to find out which bloggers blogged on blog 89 for all or part of 2006-09-01 to 2006-09-02.

I use this dataset as an example, but we might have completely different circumstances such as Blogger 20 being replaced on the third day of the date range, being replaced by blogger 29 for 10 days and then quitting due to lack of time and the blog going unmanned for 5 days before Blogger 20 decides to step back in on day 25.

The flags are the start/finish, obviously.

Transform SubSelect In OUTER JOIN
maybe I'm simply to dump but I could not transform this SQL-Statment
which uses a Sub-select and create on that uses an OUTER JOIN ....

Ranking Student Grade? With Subquery/subselect?
I am a mySQL newbie here and have some problem defining the mySQL 4.0.14
or 3.23 SQL to get student grade ranking where tied grade have the same
rank.

I used to set it through MS Access 2002 and use this kind of query:

SELECT nilai.studentNIS, nilai.studenttestmark,
(SELECT COUNT(*) FROM tblStudentGrades
WHERE [studenttestmark]>[Nilai].[studenttestmark];)+1 AS NomorUrut FROM
tblStudentGrades AS nilai ORDER BY nilai.studenttestmark DESC;

I've been looking around mySQL documentation and read that subquery can
be redefined as INNER JOIN or using two SQL statement via variable? I
have no idea on the basics of how to set it out though.

Could one of you please help give a me a sample on how this kind of
query should be done on mySQL? Is it possible to do it in single line?
And without having to use PHP/Perl scripts?

Or maybe I should have approach it differently?