10 things in MySQL (that won’t work as expected)

#10. Searching for a NULL

1.SELECT  *
2.FROM    a
3.WHERE   a.column NULL

In SQL, a NULL is never equal to anything, even another NULL. This query won’t return anything and in fact will be thrown out by the optimizer when building the plan.

When searching for NULL values, use this instead:

1.SELECT  *
2.FROM    a
3.WHERE   a.column IS NULL

#9. LEFT JOIN with additional conditions

1.SELECT  *
2.FROM    a
3.LEFT JOIN
4.b
5.ON      b.a = a.id
6.WHERE   b.column 'something'

LEFT JOIN is like INNER JOIN except that it will return each record from a at least once, substituting missing fields from b with NULL values, if there are no actual matching records.

The WHERE condition, however, is evaluated after the LEFT JOIN so the query above checks column after it had been joined. And as we learned earlier, no NULL value can satisfy an equality condition, so the records from a without corresponding record from b will unavoidably be filtered out.

Essentially, this query is an INNER JOIN, only less efficient.

To match only the records with b.column = 'something' (while still returning all records from a), this condition should be moved into ON clause:

1.SELECT  *
2.FROM    a
3.LEFT JOIN
4.b
5.ON      b.a = a.id
6.AND b.column 'something'

#8. Less than a value but not a NULL

Quite often I see the queries like this:

1.SELECT  *
2.FROM    b
3.WHERE   b.column 'something'
4.AND b.column IS NOT NULL

This is actually not an error: this query is valid and will do what’s intended. However, IS NOT NULL here is redundant.

If b.column is a NULL, then b.column < 'something' will never be satisfied, since any comparison to NULLevaluates to a boolean NULL and does not pass the filter.

It is interesting that this additional NULL check is never used for greater than queries (like in b.column > 'something').

This is because NULL go first in ORDER BY in MySQL and hence are incorrectly considered less than any other value by some people.

This query can be simplified:

1.SELECT  *
2.FROM    b
3.WHERE   b.column 'something'

and will still never return a NULL in b.column.

#7. Joining on NULL

1.SELECT  *
2.FROM    a
3.JOIN    b
4.ON      a.column = b.column

When column is nullable in both tables, this query won’t return a match of two NULLs for the reasons described above: no NULLs are equal.

Here’s a query to do that:

1.SELECT  *
2.FROM    a
3.JOIN    b
4.ON      a.column = b.column
5.OR (a.column IS NULL AND b.column IS NULL)

MySQL‘s optimizer treats this as an equijoin and provides a special join condition, ref_or_null.

#6. NOT IN with NULL values

1.SELECT  a.*
2.FROM    a
3.WHERE   a.column NOT IN
4.(
5.SELECT column
6.FROM    b
7.)

This query will never return anything if there is but a single NULL in b.column. As with other predicates, bothIN and NOT IN against NULL evaluate to NULL.

This should be rewritten using a NOT EXISTS:

1.SELECT  a.*
2.FROM    a
3.WHERE   NOT EXISTS
4.(
5.SELECT NULL
6.FROM    b
7.WHERE   b.column = a.column
8.)

Unlike INEXISTS always evaluates to either true or false.

#5. Ordering random samples

1.SELECT  *
2.FROM    a
3.ORDER BY
4.RAND(), column
5.LIMIT 10

This query attempts to select 10 random records ordered by column.

ORDER BY orders the output lexicographically: that is, the records are only ordered on the second expression when the values of the first expression are equal.

However, the results of RAND() are, well, random. It’s infeasible that the values of RAND() will match, so ordering on column after RAND() is quite useless.

To order the randomly sampled records, use this query:

01.SELECT  *
02.FROM    (
03.SELECT  *
04.FROM    mytable
05.ORDER BY
06.RAND()
07.LIMIT 10
08.) q
09.ORDER BY
10.column

#4. Sampling arbitrary record from a group

This query intends to select one column from each group (defined by grouper)

1.SELECT  DISTINCT(grouper), a.*
2.FROM    a

DISTINCT is not a function, it’s a part of SELECT clause. It applies to all columns in the SELECT list, and the parentheses here may just be omitted. This query may and will select the duplicates on grouper (if the values in at least one of the other columns differ).

Sometimes, it’s worked around using this query (which relies on MySQL‘s extensions to GROUP BY):

1.SELECT  a.*
2.FROM    a
3.GROUP BY
4.grouper

Unaggregated columns returned within each group are arbitrarily taken.

At first, this appears to be a nice solution, but it has quite a serious drawback. It relies on the assumption that all values returned, though taken arbitrarily from the group, will still belong to one record.

Though with current implementation is seems to be so, it’s not documented and can be changed in any moment (especially if MySQL will ever learn to apply index_union after GROUP BY). So it’s not safe to rely on this behavior.

This query would be easy to rewrite in a cleaner way if MySQL supported analytic functions. However, it’s still possible to make do without them, if the table has a PRIMARY KEY defined:

01.SELECT  a.*
02.FROM    (
03.SELECT  DISTINCT grouper
04.FROM    a
05.) ao
06.JOIN    a
07.ON      a.id =
08.(
09.SELECT  id
10.FROM    a ai
11.WHERE   ai.grouper = ao.grouper
12.LIMIT 1
13.)

#3. Sampling first record from a group

This is a variation of the previous query:

1.SELECT  a.*
2.FROM    a
3.GROUP BY
4.grouper
5.ORDER BY
6.MIN(id) DESC

Unlike the previous query, this one attempts to select the record holding the minimal id.

Again: it is not guaranteed that the unaggregated values returned by a.* will belong to a record holdingMIN(id) (or even to a single record at all).

Here’s how to do it in a clean way:

01.SELECT  a.*
02.FROM    (
03.SELECT  DISTINCT grouper
04.FROM    a
05.) ao
06.JOIN    a
07.ON      a.id =
08.(
09.SELECT  id
10.FROM    a ai
11.WHERE   ai.grouper = ao.grouper
12.ORDER BY
13.ai.grouper, ai.id
14.LIMIT 1
15.)

This query is just like the previous one but with ORDER BY added to ensure that the first record in id order will be returned.

#2. IN and comma-separated list of values

This query attempts to match the value of column against any of those provided in a comma-separated string:

1.SELECT  *
2.FROM    a
3.WHERE   column IN ('1, 2, 3')

This does not work because the string is not expanded in the IN list.

Instead, if column column is a VARCHAR, it is compared (as a string) to the whole list (also as a string), and of course will never match. If column is of a numeric type, the list is cast into the numeric type as well (and only the first item will match, at best).

The correct way to deal with this query would be rewriting it as a proper IN list

1.SELECT  *
2.FROM    a
3.WHERE   column IN (1, 2, 3)

, or as an inline view:

01.SELECT  *
02.FROM    (
03.SELECT  AS id
04.UNION ALL
05.SELECT  AS id
06.UNION ALL
07.SELECT  AS id
08.) q
09.JOIN    a
10.ON      a.column = q.id

, but this is not always possible.

To work around this without changing the query parameters, one can use FIND_IN_SET:

1.SELECT  *
2.FROM    a
3.WHERE   FIND_IN_SET(column'1,2,3')

This function, however, is not sargable and a full table scan will be performed on a.

#1. LEFT JOIN with COUNT(*)

1.SELECT  a.id, COUNT(*)
2.FROM    a
3.LEFT JOIN
4.b
5.ON      b.a = a.id
6.GROUP BY
7.a.id

This query intends to count number of matches in b for each record in a.

The problem is that COUNT(*) will never return a 0 in such a query. If there is no match for a certain record ina, the record will be still returned and counted.

COUNT should be made to count only the actual records in b. Since COUNT(*), when called with an argument, ignores NULLs, we can pass b.a to it. As a join key, it can never be a null in an actual match, but will be if there were no match:

1.SELECT  a.id, COUNT(b.a)
2.FROM    a
3.LEFT JOIN
4.b
5.ON      b.a = a.id
6.GROUP BY
7.a.id

P.S. In case you were wondering: no, the pictures don’t have any special meaning. I just liked them.

via Tumblr http://ift.tt/Yaaa48

Anuncios

Responder

Introduce tus datos o haz clic en un icono para iniciar sesión:

Logo de WordPress.com

Estás comentando usando tu cuenta de WordPress.com. Cerrar sesión / Cambiar )

Imagen de Twitter

Estás comentando usando tu cuenta de Twitter. Cerrar sesión / Cambiar )

Foto de Facebook

Estás comentando usando tu cuenta de Facebook. Cerrar sesión / Cambiar )

Google+ photo

Estás comentando usando tu cuenta de Google+. Cerrar sesión / Cambiar )

Conectando a %s