I have a sql statement I've created, and I need to transform it to use explict join operators so that all compare against constant clauses, and only compare against constant clauses, appear in the where clause for the query.
I am not sure how to make this change though, can anyone show me how I would do this? Here is what I have:
select S.sname
from P, J, S, SPJ
where P.pname = 'Bolt'
and J.city = 'London'
and P.p# = SPJ.p#
and J.j# = SPJ.j#
and S.s# = SPJ.s#;
If I understand you, you are looking to convert from sql89 syntax to an inner join.
It would look like this:
select
S.sname
from
P
inner join SPJ on `P.p#` = `SPJ.p#` and P.pname = 'Bolt'
inner join J on `SPJ.j#` = `J.j#` and J.city = 'London'
inner join S on `SPJ.s#` = `S.s#`
I have added the pname and city restrictions to the join syntax because that appears to be what you asked for. These can be left in the where clause as well however.
Also note that extended or special characters in column names in mysql (like p#) must be enclosed in backticks.
You want something like this:
SELECT S.sname
FROM P INNER JOIN SPJ ON P.p#=SPJ.p#
INNER JOIN J ON J.j# = SPJ.j#
INNER JOIN S ON S.s# = SPJ.s#
WHERE P.pname = 'Bolt'
AND J.city = 'London';
The conditions that are used to combine tables are placed in the JOIN clauses, and the other conditions are left in the WHERE clause.
Related
I have the following MySQL query:
SELECT p.*,
IF(COUNT(ms.PropertyID) > 0,1,0) AS Contacted,
pm.MediaID,
date_format(p.AvailableFrom, '%d %b %Y') AS 'AvailableFrom',
astext(pg.Geometry) AS Geometry
FROM property p, propertygeometry pg
JOIN shortlist sl ON sl.PropertyID = p.id AND sl.MemberID = 384216
LEFT JOIN message ms ON ms.PropertyID = p.id AND ms.SenderID = 384216
LEFT JOIN property_media pm ON pm.PropertyID = p.id AND pm.IsPrimary = 1
WHERE p.paused = 0
AND p.PropertyGeometryID = pg.id
GROUP BY p.id
And I'm getting this error:
#1054 - Unknown column 'p.id' in 'on clause'
As far as I can see the query looks right, any idea what could be wrong?
Don't mix ANSI-89 style and ANSI-92 style joins. They have different precedence which can lead to confusing errors, and that is what has happened here. Your query is being interpreted as follows:
FROM property p, (
propertygeometry pg
JOIN shortlist sl ON sl.PropertyID = p.id AND sl.MemberID = 384216
...
)
In the above, the joins using the JOIN keyword are evaluated first before the comma-style join is even considered. At that point the table p isn't yet declared.
From the MySQL manual:
However, the precedence of the comma operator is less than of INNER JOIN, CROSS JOIN, LEFT JOIN, and so on. If you mix comma joins with the other join types when there is a join condition, an error of the form Unknown column 'col_name' in 'on clause' may occur. Information about dealing with this problem is given later in this section.
I'd recommend always using ANSI-92 style joins, i.e. using the JOIN keyword:
SELECT p.*,
IF(COUNT(ms.PropertyID) > 0,1,0) AS Contacted,
pm.MediaID,
date_format(p.AvailableFrom, '%d %b %Y') AS 'AvailableFrom',
astext(pg.Geometry) AS Geometry
FROM property p
JOIN propertygeometry pg ON p.PropertyGeometryID = pg.id
JOIN shortlist sl ON sl.PropertyID = p.id AND sl.MemberID = 384216
LEFT JOIN message ms ON ms.PropertyID = p.id AND ms.SenderID = 384216
LEFT JOIN property_media pm ON pm.PropertyID = p.id AND pm.IsPrimary = 1
WHERE p.paused = 0
GROUP BY p.id
Related:
Why isn't SQL ANSI-92 standard better adopted over ANSI-89?
As stated before there is a precedence issue using joins via the comma operator where the LEFT JOIN will be executed and so references to table aliases won't exist at that time. Though you can implicitly tell MySQL to use a JOIN via that statement you may also tell MySQL to evaluate the comma joined tables first, then execute left join thusly:
SELECT p.*,
IF(COUNT(ms.PropertyID) > 0,1,0) AS Contacted,
pm.MediaID,
date_format(p.AvailableFrom, '%d %b %Y') AS 'AvailableFrom',
astext(pg.Geometry) AS Geometry
FROM (property p, propertygeometry pg)
JOIN shortlist sl ON sl.PropertyID = p.id AND sl.MemberID = 384216
LEFT JOIN message ms ON ms.PropertyID = p.id AND ms.SenderID = 384216
LEFT JOIN property_media pm ON pm.PropertyID = p.id AND pm.IsPrimary = 1
WHERE p.paused = 0
AND p.PropertyGeometryID = pg.id
GROUP BY p.id
Notice the comma separated tables are contained within parenthesis (). The table aliases and columns will now be available to your other JOINs.
I bumped into this error unknown column, the diff is the query is built thru HQL inside session.executeQuery("select id, name, sum(paid), custType from cust group by brand") that's why having to manually type inner join or join keyword is not an option as the hql is the one generating it.
it produces a query sumthing like this:
select cust_id, name, sum(paid), c.custTypeId
from customer c, custType ct
on c.custTypeId = ct.custTypeId
it says "unknown c.custTypeId" column when I am 101% sure it bears that column.
My classes/relations:
Customer {
Integer custId
CustomerType custType
}
CustomerType{
Integer custTypeId
string code
}
the problem lies in the comma in "from customer, custType" line. it should be with the word JOIN as the answer stated above. but since it is HQL and is being generated, I can't do that. What I did is modified by query and instead of typing select custType, I typed select custType.id, custType.code
I know it's basic but for first timers like me, it was a struggle.
If this helps someone (and a note to future myself), I was getting this error when trying to execute the following queries in MariaDB:
SELECT a.name, b.name
FROM `cities` as a
INNER JOIN `countries` as b
ON `a.country_id` = `b`.`id`;
whereas I should have written it like:
SELECT a.name, b.name
FROM `cities` as a
INNER JOIN `countries` as b
ON `a`.`country_id` = `b`.`id`;
I'll leave it to the reader to spot the difference as an exercise. :)
I'm trying to study SQL.
I have a problem with JOIN
I want to display ref_id, pro_name, class_name but I couldn't.
I find EFFICIENT solution.
MY QUERY (DOESN'T WORK)
SELECT
ref_id, pro_name, class_name
FROM
RC, RP, PP, LP
WHERE
RC.ref_id = RP.ref_id
Avoid using commas be CROSS JOIN
You could use JOIN to instead of commas
like this.
SELECT
RP.ref_id, PP.pro_name, LP.class_name
FROM
RP
LEFT JOIN RC ON RC.ref_id = RP.ref_id
LEFT JOIN PP ON PP.pro_id = RP.pro_id
LEFT JOIN LP ON LP.lec_id = RP.lec_id
Never use commas in the FROM clause. Always use proper, explicit, standard JOIN syntax.
You would seem to want:
select rp.pro_id, pp.pro_name, lp.class_name
from rp left join
pp
on rp.pro_id = pp.pro_id left join
lp
on rp.lec_id = lp.lec_id;
Note the use of left join. This ensure that all rows are in the result set, even when one or the other joins doesn't find a matching record.
From what I can see, the table rc is not needed to answer this specific question.
I have the following query to return all user_attributes and attributes which have a specified tag:
SELECT `user_attributes`.*, `attributes`.*
FROM `user_attributes`
INNER JOIN `attributes` ON (`attributes`.`id` = `user_attributes`.`attribute_id`)
INNER JOIN `user_tags` ON (`attributes`.`id` = `user_tags`.`attribute_id`)
INNER JOIN `tags` ON (`user_tags`.`tag_id` = `tags`.`id`)
WHERE `user_attributes`.`user_id` = '1'
AND `tags`.`title` IN ('tag1')
I would like to adjust the query so that it finds all values that have 2 tags. At the moment I have:
SELECT `user_attributes`.*
FROM `user_attributes`
INNER JOIN `attributes` ON (`attributes`.`id` = `user_attributes`.`attribute_id`)
INNER JOIN `user_tags` ON (`attributes`.`id` = `user_tags`.`attribute_id`)
INNER JOIN `tags` ON (`user_tags`.`tag_id` = `tags`.`id`)
WHERE `user_attributes`.`user_id` = '1'
AND `tags`.`title` IN ('tag1', 'tag2')
HAVING (COUNT(DISTINCT `tags`.`title`) = 2)
Is it breaking because I'm using HAVING without a GROUP BY?
HAVING should be used in combination with GROUP BY indeed. MySQL is the only database what will handle HAVING without GROUP BY as some kind off WHERE
Some more proof for the downvoter..
MySQL http://www.sqlfiddle.com/#!2/ba8d6/3 (this is WRONG and looks like HAVING IS USED AS WHERE)
Oracle http://www.sqlfiddle.com/#!4/ba8d6/1 (this is correct ORA-00979: not a GROUP BY expression Oracle is missing the GROUP BY)
Postgresql http://www.sqlfiddle.com/#!1/ba8d6/2 (this is correct ERROR: column "data.username" must appear in the GROUP BY clause or be used in an aggregate function Postgresql wants to have an GROUP BY
I have this query:
SELECT hit.timestamp,hit.id,config.Name,hit.meter_id,levels.LevelName, pos.sm_pos , hit.hit_value
FROM pos,hit,controllers,levels,config
WHERE hit.id=config.id
AND hit.meter_id=levels.id
AND pos.id=hit.id
AND pos.controller_id=controllers.id;
How to make an inner join query from this? With aliases or something? I was looking and I can't find anything for multiple table query.
The way you are joining tables is outdated now. It was used eartlier, now we use keyword like INNER/NATURAL/LEFT OUTER/RIGHT OUTER/CROSS etc. to join tables on basis of requirement.
Refer Join in Mysql
SELECT hit.timestamp,
hit.id,
config.Name,
hit.meter_id,
levels.LevelName,
pos.sm_pos,
hit.hit_value
FROM hit
INNER JOIN config
ON hit.id = config.id
INNER JOIN levels
ON hit.meter_id = levels.id
INNER JOIN POS
ON pos.id = hit.id
INNER JOIN controllers
ON pos.controller_id = controllers.id;
Note : The query posted by you is according the SQL-89 standard and the second posted by me is according to SQL-92.
The SQL-92 standard introduced INNER JOIN .. ON and OUTER JOIN .. ON in order to replace the more complex(?) syntax of SQL-89.
If you want to reformat your query, you can do it like this:
SELECT
H.timestamp,
H.id,
F.Name,
H.meter_id,
L.LevelName,
P.sm_pos,
H.hit_value
FROM pos AS P
INNER JOIN controllers AS C ON P.controller_id = C.id
INNER JOIN hit AS H ON P.id = H.id
INNER JOIN levels AS L ON H.meter_id = L.id
INNER JOIN config AS F ON H.id =F.id;
Notice that I've taken the liberty to add aliases on your table names, this can simplify your queries alot.
To understand how joins work in MySQL, read about it here in the manual. A good tutorial about joins was written by Jeff Atwood, you can find it here.
Could anybody help me?
With this query I am getting the ids, but it is not putting the separators when subscriber_data.fieldid is null. For example instead of coming 2,,12 it comes 2,12 when the value for 4 is null...
I think the problem is on the Join with subquery, but i couldn't make it with two left joins in the main query also...
This is the query im using:
SELECT
list_subscribers.emailaddress,
(SELECT
GROUP_CONCAT(IFNULL(customfields.fieldid,'') SEPARATOR '","')
FROM customfields
LEFT JOIN subscribers_data
ON subscribers_data.fieldid = customfields.fieldid
WHERE
customfields.fieldid IN (2,4,12,13,14,15,17,19,20,21,22,23,16,26,27)
AND
list_subscribers.subscriberid = subscribers_data.subscriberid
) AS data FROM list_subscribers
Thanks everyone.
The left join is useless. Because you have a condition on subscriber_data in the WHERE clause, that subquery will not return those rows for which there is no matching subscriber_data, so it effectively works as if you used INNER JOIN. You should add that condition to the left join condition, but it is impossible in this query layout. Values from the outer query are not allowed in join conditions in the inner query.
You could change it, but apparently you need to join three tables, where the middle table, subscriber_data, that links them all together, is optional. That doesn't really make sense.
Or maybe customfields is the table that is optional, but in that case, you should have reversed the table or used a RIGHT JOIN.
In conclusion, I think you meant to write this:
select
s.emailaddress,
GROUPCONCAT(IFNULL(f.fieldid, '') SEPARATOR '","')
from
list_subscribers s
inner join subscribers_data d on d.subscriberid = s.subscriberid
left join customfields f on f.fieldid = d.fieldid
where
d.fieldid in (2,4,12,13,14,15,17,19,20,21,22,23,16,26,27)
group by
s.emailaddress
Or do you want to list the id's of the fields that are filled for the subscriber(s)? In that case, it would be:
select
s.emailaddress,
GROUPCONCAT(IFNULL(d.fieldid, '') SEPARATOR '","')
from
list_subscribers s
cross join customfields f
left join subscribers_data d on
d.subscriberid = s.subscriberid and
d.fieldid = f.fieldid
where
f.fieldid in (2,4,12,13,14,15,17,19,20,21,22,23,16,26,27)
group by
s.emailaddress