I'am trying to understand what causes the following, maybe you could help me:
I have a query like:
select field1,fieldDate from table1
union all
select field1,fieldDate from table2
order by fieldDate desc
and the another one like this:
select field1,field2,fieldDate from table1
union all
select field1,field2,fieldDate from table2
order by fieldDate desc
So basically they are the same with the exception that in the second I retrieve an extra field.
Now, both results come with a diferent ordering, but just for the cases that the dates are exacly the same. For example there are 2 rows (row1,row2) with date 2009-11-25 09:41:55. For query 1 row1 comes before row2 and for query 2 row2 comes before row1.
Does somebody knows why this happens?
Thanks,
Regards
The ordering based on any fields that you don't explicitly order by is undefined, and the optimizer can change the ordering if it thinks that results in a better execution plan. Given two rows with the exact same value in the order by field you can not depend on them being in any particularly order in relation to each other unless you explicitly order by another field with different values.
Can you do this
select * from ( select
field1,field2,fieldDate, 0 as ordercol from table1
union all select
field1,field2,fieldDate, 1 as ordercol from table2) t1
order by fieldDate desc, ordercol asc
Straight from the MySQl manual, to user order by on a union you have to parenthesis the individual tables.
(select field1,fieldDate from table1)
union all
(select field1,fieldDate from table2)
order by fieldDate desc
This is not SQL standards compliant! The code you entered should order the union of both tables but to my surprise MySQL has the above syntax.
The order in which rows with the same fieldDate are returned can differ for each query execution. Usually this order will be the same but you should not count on it. If you want any extra ordering state more order by fields.
EDIT: This answer is wrong: the order by works on the entire union. I'll leave it here to save others the trouble :)
Your order by only works on the second part of the union. You can use a subquery to make the order by work on the entire union:
select field1,field2,fieldDate
from (
select field1,field2,fieldDate
from table1
union all
select field1,field2,fieldDate
from table2
) SubQueryName
order by fieldDate desc
Related
I am trying to get my head around using GROUP_CONCAT within MYSQL.
Basically I have the following table, table1:
id, field1, field2, active
I want to bring back 5 rows within the table but in random order. So I'm using this:
SELECT GROUP_CONCAT(id ORDER BY rand()) FROM table1 WHERE active=1
This behaves as I would expect. I then want to use the output to select the other columns (field1, field2) from the table and display the results.
So I've tried using:
SELECT *
FROM table1
WHERE id IN
(
SELECT GROUP_CONCAT(id ORDER BY rand()) as id FROM table1 WHERE active=1
);
I expected something like the above to work but I cant figure out why it doesn't. It DOES bring back results but not all of them, (i.e.) my table contains 10 rows. 6 rows are set to active=1. Therefore I would expect 6 rows to be returned ... this isn't happening I may get 1,2 or 0.
Additionally if it helps I'd like to limit the number of results returned by the sub-query to 3 but adding LIMIT doesn't seem to have any affect on the results returned.
Thank you in advance for your help
I think this is what you are looking for. This will bring back 5 random active rows.
SELECT *
FROM table1
WHERE active=1
ORDER BY RAND()
LIMIT 5;
why not use this :
SELECT *, GROUP_CONCAT(id ORDER BY rand()) as randoms FROM table1 WHERE active=1
If I understand correctly, you are trying to build a query like this:
select *
from table1
where id in (1,2,3,4,5) -- Just an example
and you are trying to "fill" the in condition with a group_concat() result.
That's not the way to do it.
You only need to specify the subquery in the parenthesis:
select *
from table1
where id in (select id from table1 where active=1)
Notice some additional things:
The order by rand() is irrelevant, because the in () will be evaluated regardless of the order of the values.
In this particular scenario, I would recommend to use a join instead of in.
Using join:
select t1.*
from
table1 as t1
inner join table1 as t2 on t1.id = t2.id
where t2.active=1
Exist a better way to do what the following SQL query does? I have the feeling that table1 will be searched twice and may be that can be avoided with some trick and increase the efficient of the query, but I just can't figure out how ;( Here is the query (in MySQL):
SELECT a, SUM(count)
FROM table1
GROUP BY a
HAVING SUM(count) = (SELECT SUM(count) as total FROM table1 GROUP BY a ORDER BY total DESC LIMIT 1)
The goal is return the number(s) with the major accumulate, with its accumulate.
being table1 a two field table like:
a,count
1,10
1,30
1,0
2,1
2,100
2,4
3,10
4,50
4,55
The result with that data sample is:
2,105
4,105
Thanks in advance.
SELECT a, total FROM
(SELECT a AS a, SUM(COUNT) AS total
FROM table1
GROUP BY a) AS xyz
HAVING total = MAX(total)
Hope this will work for you
This sub-query is executed only once, and you don't have to bother with creating any pre-query as other answers may suggest (although doing so this is still correct, just not needed). Database engine will realise, that the sub-query is not using any variable dependent on the other part of the query. You can use EXPLAIN to see how the query is executed.
More on the topic in this answer:
https://stackoverflow.com/a/658954/1821029
I think you could probably do it by moving your HAVING sub-select query into its on prequery. Since it will always include a single row, you won't require any "JOIN", and it does not have to keep recomputing the COUNT(*) every time the HAVING is applied. Do it once, then the rest
SELECT
a,
SUM(count)
FROM
table1,
( SELECT SUM(count) as total
FROM table1
GROUP BY a
ORDER BY total DESC
LIMIT 1 ) PreQuery
GROUP BY
a
HAVING
SUM(count) = PreQuery.Total
This query return one row with two columns:
1- a list of comma separated values of "a" column, which have the biggest "Total"
2- and the biggest Total value
select group_concat(a), Total
from
(select a, sum(count) as Total
from table1
group by a) OnTableQuery
group by Total
order by Total desc
limit 1
Note that it queries table1 just one time. The query was already tested.
Assuming table1 and table2 both have a large number of rows (ie several hundred thousand), is the following an inefficient query?
Edit: Order by field added.
SELECT * FROM (
SELECT title, updated FROM table1
UNION
SELECT title, updated FROM table2
) AS query
ORDER BY updated DESC
LIMIT 25
If you absolutely need distinct results, another possibility is to use union all and a group by clause instead:
SELECT title FROM (
SELECT title FROM table1 group by title
UNION ALL
SELECT title FROM table2 group by title
) AS query
group by title
LIMIT 25;
Testing this without the limit clause on an indexed ID column from two tables with ~920K rows each in a test database (at $work) resulted in a bit over a second with the query above and about 17 seconds via a union.
this should be even faster - but then I see no ORDER BY so what 25 records do you actually want?
SELECT * FROM (
SELECT title FROM table1 LIMIT 25
UNION
SELECT title FROM table2 LIMIT 25
) AS query
LIMIT 25
UNION must make an extra pass to fetch the distinct records, so you should use UNION ALL.
Yes, use order by and limits in the inner queries.
SELECT * FROM (
(SELECT title FROM table1 ORDER BY title ASC LIMIT C)
UNION
(SELECT title FROM table2 ORDER BY title ASC LIMIT C)
) AS query
LIMIT 25
This will only go through C rows instead of N (hundreds of thousands). The ORDER BY is necessary and should be on an indexed column.
C is a heuristic constant that should be tuned according to the domain. If you only expect a few duplicates, C=50-100 is probably ok.
You can also find out this for yourself by using EXPLAIN.
SELECT
....
GROUP BY
c.city_id
ORDER BY p.name desc
UNION
SELECT
...
GROUP BY
c.city_id, p
ERROR 1221 (HY000): Incorrect usage of UNION and ORDER BY
is there a way to have this format because i want the top query to have an orderby do i need to have the same orderby on the bottom query
The ORDER By comes at the end
select *
from..
union all
select *
from...
order by....
what you can do if you want the first query to show up first is this
select *, 1 as SortOrder
from..
union all
select * ,2 as SortOrder
from...
order by SortOrder,<other columns>...
ORDER BY should be at the end of your select statement, not before the UNION.
See here and here for more information on UNION syntax.
You can't use an order by on the select queries that will be joined by the UNION. You can, if you want, select everything afterwards and add an order by then.
In standard SQL, the ORDER BY comes at the end of the UNION'd queries and is applied to the final result from those queries.
But...
MySQL allows you to use an ORDER BY within a UNION statement if you enclose it in brackets:
( SELECT ....
FROM ...
GROUP BY c.city_id
ORDER BY p.name DESC )
UNION
SELECT ...
FROM ...
GROUP BY c.city_id
...which'll also allow you to use LIMIT...
As the other answers say, It is being parsed as
SELECT { unordered_stuff UNION SELECT unordered_stuff } ORDER BY foo
Some, though I don't believe all, databases support an alternate disambigiouation syntax. This is from the Pg lists.
(SELECT * from foo where priority = 1 order by seniority)
UNION ALL
(select * from foo where priority > 1 order by seniority, priority)
Otherwise the ORDER BY is considered to apply to the whole UNION result
(it's effectively got lower binding priority than the UNION). Note also
that you *must* use UNION ALL, else UNION will attempt to eliminate
duplicates, and mess up the sort order while at it.
See also Bruno's solution nearby. Not sure which of these approaches
would be faster; try both.
I have a few tables that have similar fields but not exactly the same.
The same fields they have are description (text field) and modified (unixtime)
I would like to select the last modified items from these tables based on unixtime. I cant use UNION since the tables aren't the same and the multiple table select times out.
I've been trying to look for this but no luck, either people are using JOINS or SELECT A., B. FROM table A, table B
How different they are?
Maybe you can get the common fields out:
select t1.name1 as name from table1
union
select t2.name2 as name from table2
Try this:
SELECT
IF (A.modified > B.modified, A.modified, B.modified) AS modified,
IF (A.modified > B.modified, A.description, B.description) AS description,
FROM
(SELECT description, modified FROM A ORDER BY modified DESC LIMIT 1) AS A,
(SELECT description, modified FROM B ORDER BY modified DESC LIMIT 1) AS B
LIMIT 1
However, it's pretty much the same as just doing two queries (only more complicated) so I wouldn't recommend it.
Try adding desc index on 'modified' if your select timesou, and use limit on select to return just one (last) row.
Then you can:
SELECT
A,B,C,D, desc, modified
FROM
TABLEA
UNION ALL
SELECT
CAST(E as <A type>), CAST(F AS <B type>) ..., desc, modified
FROM
TABLE B