I am trying to use MySQL benchmark to test some queries. But, I am running to an error.
SELECT benchmark (10000, (select title from user));
and in return I get this error;
ERROR 1242 (21000): Subquery returns more than 1 row
Does anyone know how to benchmark a query?
Thanks
select title from user
This returns multiple rows, which won't work.
Refer to this link:
http://dev.mysql.com/doc/refman/5.0/en/information-functions.html#function_benchmark
The expression you pass must return a scalar result.
You need to change the query such that it returns a single row:
ex:
select title from user where user_name = 'some_user'
you can use the mysqlslap utility to benchmark queries, see:
http://dev.mysql.com/doc/refman/5.1/en/mysqlslap.html
From http://dev.mysql.com/doc/refman/5.0/en/information-functions.html#function_benchmark
Only scalar expressions can be used.
Although the expression can be a
subquery, it must return a single
column and at most a single row. For
example, BENCHMARK(10, (SELECT * FROM
t)) will fail if the table t has more
than one column or more than one row.
Try
SELECT BENCHMARK(10000, (SELECT title FROM user LIMIT 1));
Related
I have a PHP array with numbers of ID's in it. These numbers are already ordered.
Now i would like to get my result via the IN() method, to get all of the ID's.
However, these ID's should be ordered like in the IN method.
For example:
IN(4,7,3,8,9)
Should give a result like:
4 - Article 4
7 - Article 7
3 - Article 3
8 - Article 8
9 - Article 9
Any suggestions? Maybe there is a function to do this?
Thanks!
I think you may be looking for function FIELD -- while normally thought of as a string function, it works fine for numbers, too!
ORDER BY FIELD(field_name, 3,2,5,7,8,1)
You could use FIELD():
ORDER BY FIELD(id, 3,2,5,7,8,1)
Returns the index (position) of str in the str1, str2, str3, ... list. Returns 0 if str is not found.
It's kind of an ugly hack though, so really only use it if you have no other choice. Sorting the output in your app may be better.
Standard SQL does not provide a way to do this (MySQL may, but I prefer solutions that are vendor-neutral so I can switch DBMS' at any time).
This is something you should do in post-processing after the result set is returned. SQL can only return them in an order specified in the "order by" clause (or in any order if there's no such clause).
The other possibility (though I don't like it, I'm honor-bound to give you the choice) is to make multiple trips to the database, one for each ID, and process them as they come in:
select * from tbl where article_id = 4;
// Process those.
select * from tbl where article_id = 7;
// Process those.
: : : : :
select * from tbl where article_id = 9;
// Process those.
You'll just need to give the correct order by statement.
SELECT ID FROM myTable WHERE ID IN(1,2,3,4) ORDER BY ID
Why would you want to get your data ordered unordered like in your example?
If you don't mind concatening long queries, try that way:
SELECT ID FROM myTable WHERE ID=1
UNION
SELECT ID FROM myTable WHERE ID=3
UNION
SELECT ID FROM myTable WHERE ID=2
SQL Queries /P1/
SELECT EXISTS(SELECT /p2/ FROM table WHERE id = 1)
SELECT /p2/ FROM table WHERE id = 1 LIMIT 1
SQL SELECT /P2/
COUNT(id)
id
PHP PDO Function /P3/
fetchColumn()
rowCount()
From the following 3 Parts, What is the best method to check if a row exists or not with and without the ability to retrieve data like.
Retrievable:
/Query/ SELECT id FROM table WHERE id = 1 LIMIT 1
/Function/ rowCount()
Irretrievable
/Query/ SELECT EXISTS(SELECT COUNT(id) FROM table WHERE id = 1)
/Function/ fetchColumn()
In your opinion, What is the best way to do that?
By best I guess you mean consuming the least resources on both MySQL server and client.
That is this:
SELECT COUNT(*) count FROM table WHERE id=1
You get a one-row, one-column result set. If that column is zero, the row was not found. If the column is one, a row was found. If the column is greater that one, multiple rows were found.
This is a good solution for a few reasons.
COUNT(*) is decently efficient, especially if id is indexed.
It has a simple code path in your client software, because it always returns just one row. You don't have to sweat edge cases like no rows or multiple rows.
The SQL is as clear as it can be about what you're trying to do. That's helpful to the next person to work on your code.
Adding LIMIT 1 will do nothing if added to this query. It is already a one-row result set, inherently. You can add it, but then you'll make the next person looking at your code wonder what you were trying to do, and wonder whether you made some kind of mistake.
COUNT(*) counts all rows that match the WHERE statement. COUNT(id) is slightly slower because it counts all rows unless their id values are null. It has to make that check. For that reason, people usually use COUNT(*) unless there's some chance they want to ignore null values. If you put COUNT(id) in your code, the next person to work on it will have to spend some time figuring out whether you meant anything special by counting id rather than *.
You can use either; they give the same result.
I would like get the id and the number of id. So, I write this command sql:
SELECT count(id), id
FROM tblExample
It's doesn't work. Have you a solution for me ? For to get the value of my id and the number of id.
Or a function PHP for count my resultset.
Just add a GROUP BY id:
SELECT id, COUNT(id)
FROM tblExample
GROUP BY id;
Demo
Update:
The query you posted:
SELECT count(id), id
FROM tblExample;
Won't work in most of the RDBMS and it shouldn't. In SQL Server, you will got an error; saying that:
Column 'id' is invalid in the select list because it is not contained in either an
aggregate function or the GROUP BY clause.
Strangely though, MySQL allow this(The OP didn't say what RDBMS he is using), and in this case, it will get an arbitrary value (this is also depends on an option to set), for the id column, and the COUNT in this case would be all the id's count.
But it is not recommended to do so.
I want to do a SELECT on an empty table, but i still want to get a single record back with all the column names. I know there are other ways to get the column names from a table, but i want to know if it's possible with some sort of SELECT query.
I know this one works when i run it directly in MySQL:
SELECT * FROM cf_pagetree_elements WHERE 1=0;
But i'm using PHP + PDO (FETCH_CLASS). This just gives me an empty object back instead of an row with all the column names (with empty values). So for some reason that query doesn't work with PDO FETCH_CLASS.
$stmt = $this->db->prepare ( $sql );
$stmt->execute ( $bindings );
$result = $stmt->fetchAll ( \PDO::FETCH_CLASS, $class );
print_r($result); // Empty object... I need an object with column names
Anyone any idea if there's another method that i can try?
Adding on to what w00 answered, there's a solution that doesn't even need a dummy table
SELECT tbl.*
FROM (SELECT 1) AS ignore_me
LEFT JOIN your_table AS tbl
ON 1 = 1
LIMIT 1
In MySQL you can change WHERE 1 = 1 to just WHERE 1
To the other answers who posted about SHOW COLUMNS and the information scheme.
The OP clearly said: "I know there are other ways to get the column names from a table, but i want to know if it's possible with some sort of SELECT query."
Learn to read.
Anyway, to answer your question; No you can't. You cannot select a row from an empty table. Not even a row with empty values, from an empty table.
There is however a trick you can apply to do this.
Create an additional table called 'dummy' with just one column and one row in it:
Table: dummy
dummy_id: 1
That's all. Now you can do a select statement like this:
SELECT * FROM dummy LEFT OUTER JOIN your_table ON 1=1
This will always return one row. It does however contain the 'dummy_id' column too. You can however just ignore that ofcourse and do with the (empty) data what ever you like.
So again, this is just a trick to do it with a SELECT statement. There's no default way to get this done.
SHOW COLUMNS FROM cf_pagetree_elements;
This will give a result set explaining the table structure. You can quite easily parse the result with PHP.
Another method is to query the infomrmation schema table:
SELECT column_name FROM information_schema.columns WHERE table_name='cf_pagetree_elements';
Not really recommended though!
You could try:
SELECT * FROM information_schema.columns
WHERE table_name = "cf_pagetree_elements"
Not sure about your specific PHP+PDO approach (there may be complications), but that's the standard way to fetch column headings (field names).
this will list the columns of ANY query for PDO drivers that support getColumMeta. I am using this with SQL server and works fine even on very complex queries with aliased tables, sub-queries and unions. Gives me columns even when results are zero
<?php
// just an example of an empty query.
$query =$PDOdb->query("SELECT * from something where 1=0; ");
for ($i=0; $i<$query->columnCount(); $i++) {
echo $query->getColumnMeta($i)['name']."<br />";
}
?>
Even without PDO in the way, the database won't return the structure without at least one row. You could do this and ignore the data row:
SELECT * FROM cf_pagetree_elements LIMIT 1;
Or you could simply
DESC cf_pagetree_elements;
and deal with one row per field.
WHERE 1=0 does not work for me. It always returns empty set.
The latest PDO for SQLSVR definitely works with get column meta.
Simply set up your statement and use this to get an array of useful information:
$stmt->execute();
$meta= array();
foreach(range(0, $stmt->columnCount() - 1) as $column_index)
{
array_push($meta,$stmt->getColumnMeta($column_index));
}
Complete solution for Oracle or MySQL
for any or some columns (my goal is to get arbitrary columns exactly as they are in DB regardless of case)
for any table (w or w/o rows)
$qr = <<<SQL
SELECT $cols
FROM (SELECT NULL FROM DUAL)
LEFT JOIN $able t ON 1 = 0
SQL;
$columns = array_keys($con->query($qr)->fetchAll(PDO::FETCH_ASSOC)[0]);
if($cols === "*") {
array_shift($columns);
}
YOu could use MetaData with;
$cols = mysql_query("SHOW COLUMNS FROM $tableName", $conn);
This query runs successfully in pgAdmin, but when I transfer it in zend I get:
ERROR: subquery must return only one column...
Can someone distinguish the problem?
SELECT "trail_history"."new_audit_trail".*,
(SELECT "admin".list_category_values_new."values"
FROM "admin".list_category_values_new
WHERE CAST(seq AS character varying) = "trail_history"."new_audit_trail"."current_value"
AND "name" = "trail_history"."new_audit_trail"."name") as "values"
FROM "trail_history"."new_audit_trail"
WHERE (capno LIKE '12101062411001%')
AND (recon = '0')
ORDER BY "date_happened" DESC
Your sub-select SELECT "admin".list_category_values_new."values"... has nothing that prevents it from returning multiple rows. You need to use TOP 1 or MAX or something to ensure that only one record comes out of the sub-select.
You can correlate the sub-query so that each record in your main select gets a different single value, but if you're going to use a sub-select it can only return one row per row of output in your main select.