UPDATEs with multiple aggregate functions

[Steven An]

Howdy,

I need to write an update query with multiple aggregate functions.
Here is an example:

UPDATE t
SET
t.a = ( select avg(f.q) from dbo.foo f where f.p = t.y ),
t.b = ( select sum(f.q) from dbo.foo f where f.p = t.y )
FROM dbo.test t

Basically I need to get some aggregate statistics about the rows of
foo and store them in rows of t. The above statement works fine...but
note how the two subSelect's have the exact same WHERE clause. This
screams at me to combine them...but how? I would like to have
something like this in my query:

SELECT avg(f.q), sum(f.q) FROM dbo.foo f WHERE f.p = 2

...and somehow store the results in t.a and t.b. Is there any way to
do this?

Thanks before hand!

[John Bell]

Re: UPDATEs with multiple aggregate functions

Hi

You could try:

UPDATE t
SET a = d.a
b = d.b
FROM dbo.test t JOIN
( SELECT p,avg(q),sum(q) FROM dbo.foo GROUP BY p ) f ON f.p = t.y

John

"Steven An" <stevenan@uclin k.berkeley.edu> wrote in message
news:4953656d.0 406101610.270ac e29@posting.goo gle.com...[color=blue]
> Howdy,
>
> I need to write an update query with multiple aggregate functions.
> Here is an example:
>
> UPDATE t
> SET
> t.a = ( select avg(f.q) from dbo.foo f where f.p = t.y ),
> t.b = ( select sum(f.q) from dbo.foo f where f.p = t.y )
> FROM dbo.test t
>
> Basically I need to get some aggregate statistics about the rows of
> foo and store them in rows of t. The above statement works fine...but
> note how the two subSelect's have the exact same WHERE clause. This
> screams at me to combine them...but how? I would like to have
> something like this in my query:
>
> SELECT avg(f.q), sum(f.q) FROM dbo.foo f WHERE f.p = 2
>
> ..and somehow store the results in t.a and t.b. Is there any way to
> do this?
>
> Thanks before hand![/color]

[--CELKO--]

Re: UPDATEs with multiple aggregate functions

>> I need to write an update query with multiple aggregate functions.
Here is an example: <<

You might want to learn Standard SQL. There is no FROM clause in a
Standard SQL UPDATE statement; it would make no sense. Other products
(SQL Server, Sybase and Ingres) also use the UPDATE .. FROM syntax,
but with different semantics. So it does not port, or even worse,
when you do move it, it trashes your database. Other programmers
cannot read it and maintaining it is harder. And when Microsoft
decides to change it, you will have to do a re-write. Remember the
deprecated "*=" versus "LEFT OUTER JOIN" conversions?

The correct syntax for a searched update statement is

<update statement> ::=
UPDATE <table name>
SET <set clause list>
[WHERE <search condition>]

<set clause list> ::=
<set clause> [{ , <set clause> }...]

<set clause> ::= <object column> = <update source>

<update source> ::= <value expression> | NULL | DEFAULT

<object column> ::= <column name>

The UPDATE clause simply gives the name of the base table or updatable
view to be changed.

Notice that no correlation name is allowed in the UPDATE clause; this
is to avoid some self-referencing problems that could occur. But it
also follows the data model in Standard SQL. When you give a table
expression a correlation name, it is to act as if a materialized table
with that correlation name has been created in the database. That
table then is dropped at the end of the statement. If you allowed
correlation names in the UPDATE clause, you would be updating the
materialized table, which would then disappear and leave the base
table untouched.

The SET clause is a list of columns to be changed or made; the WHERE
clause tells the statement which rows to use. For this discussion, we
will assume the user doing the update has applicable UPDATE privileges
for each <object column>.

* The WHERE Clause

As mentioned, the most important thing to remember about the WHERE
clause is that it is optional. If there is no WHERE clause, all rows
in the table are changed. This is a common error; if you make it,
immediately execute a ROLLBACK statement.

All rows that test TRUE for the <search condition> are marked as a
subset and not as individual rows. It is also possible that this
subset will be empty. This subset is used to construct a new set of
rows that will be inserted into the table when the subset is deleted
from the table. Note that the empty subset is a valid update that
will fire declarative referential actions and triggers.

* The SET Clause

Each assignment in the <set clause list> is executed in parallel and
each SET clause changes all the qualified rows at once. Or at least
that is the theoretical model. In practice, implementations will
first mark all of the qualified rows in the table in one pass, using
the WHERE clause. If there were no problems, then the SQL engine
makes a copy of each marked row in working storage. Each SET clause
is executed based on the old row image and the results are put in the
new row image. Finally, the old rows are deleted and the new rows are
inserted. If an error occurs during all of this, then system does a
ROLLBACK, the table is left unchanged and the errors are reported.
This parallelism is not like what you find in a traditional
third-generation programming language, so it may be hard to learn.
This feature lets you write a statement that will swap the values in
two columns, thus:

UPDATE MyTable
SET a = b, b = a;

This is not the same thing as

BEGIN ATOMIC
UPDATE MyTable
SET a = b;
UPDATE MyTable
SET b = a;
END;

In the first UPDATE, columns a and b will swap values in each row. In
the second pair of UPDATEs, column a will get all of the values of
column b in each row. In the second UPDATE of the pair, a, which now
has the same value as the original value of b, will be written back
into column b -- no change at all. There are some limits as to what
the value expression can be. The same column cannot appear more than
once in a <set clause list> -- which makes sense, given the parallel
nature of the statement. Since both go into effect at the same time,
you would not know which SET clause to use.

If a subquery expression is used in a <set clause>, and it returns a
single value, the result set is cast to a scalar; if it returns an
empty, the result set is cast to a NULL; if it returns multiple rows,
a cardinality violation is raised.

When we finally get T-SQL up to SQL-92 specs, you could use row
constructors and have written:

UPDATE Test
SET ROW(a,b)
= (SELECT AVG(q), SUM(q) FROM FooWHERE Foo.p = Test.y);

[Steven An]

Re: UPDATEs with multiple aggregate functions

Just what I needed. This saves me a LOT of time :)

Many thanks!

"John Bell" <jbellnewsposts @hotmail.com> wrote in message news:<2ccyc.136 9$uW1.12016770@ news-text.cableinet. net>...[color=blue]
> Hi
>
> You could try:
>
> UPDATE t
> SET a = d.a
> b = d.b
> FROM dbo.test t JOIN
> ( SELECT p,avg(q),sum(q) FROM dbo.foo GROUP BY p ) f ON f.p = t.y
>
> John
>
> "Steven An" <stevenan@uclin k.berkeley.edu> wrote in message
> news:4953656d.0 406101610.270ac e29@posting.goo gle.com...[color=green]
> > Howdy,
> >
> > I need to write an update query with multiple aggregate functions.
> > Here is an example:
> >
> > UPDATE t
> > SET
> > t.a = ( select avg(f.q) from dbo.foo f where f.p = t.y ),
> > t.b = ( select sum(f.q) from dbo.foo f where f.p = t.y )
> > FROM dbo.test t
> >
> > Basically I need to get some aggregate statistics about the rows of
> > foo and store them in rows of t. The above statement works fine...but
> > note how the two subSelect's have the exact same WHERE clause. This
> > screams at me to combine them...but how? I would like to have
> > something like this in my query:
> >
> > SELECT avg(f.q), sum(f.q) FROM dbo.foo f WHERE f.p = 2
> >
> > ..and somehow store the results in t.a and t.b. Is there any way to
> > do this?
> >
> > Thanks before hand![/color][/color]

[Steven An]

Re: UPDATEs with multiple aggregate functions

Well, thank you for that very insightful answer. Yes, that ROW
operator is a lot like what I had in mind. I was looking for
something like SET (a,b) = (select ..). I will try that, and see
which one is faster. I have a feeling your method will be MUCH
faster, since unlike the JOIN, it does not require that I calculate
avg's and sum's for ALL rows, only the ones I want. Thank you!

jcelko212@earth link.net (--CELKO--) wrote in message news:<18c7b3c2. 0406121447.35ad ed55@posting.go ogle.com>...[color=blue][color=green][color=darkred]
> >> I need to write an update query with multiple aggregate functions.[/color][/color]
> Here is an example: <<
>
> You might want to learn Standard SQL. There is no FROM clause in a
> Standard SQL UPDATE statement; it would make no sense. Other products
> (SQL Server, Sybase and Ingres) also use the UPDATE .. FROM syntax,
> but with different semantics. So it does not port, or even worse,
> when you do move it, it trashes your database. Other programmers
> cannot read it and maintaining it is harder. And when Microsoft
> decides to change it, you will have to do a re-write. Remember the
> deprecated "*=" versus "LEFT OUTER JOIN" conversions?
>
> The correct syntax for a searched update statement is
>
> <update statement> ::=
> UPDATE <table name>
> SET <set clause list>
> [WHERE <search condition>]
>
> <set clause list> ::=
> <set clause> [{ , <set clause> }...]
>
> <set clause> ::= <object column> = <update source>
>
> <update source> ::= <value expression> | NULL | DEFAULT
>
> <object column> ::= <column name>
>
> The UPDATE clause simply gives the name of the base table or updatable
> view to be changed.
>
> Notice that no correlation name is allowed in the UPDATE clause; this
> is to avoid some self-referencing problems that could occur. But it
> also follows the data model in Standard SQL. When you give a table
> expression a correlation name, it is to act as if a materialized table
> with that correlation name has been created in the database. That
> table then is dropped at the end of the statement. If you allowed
> correlation names in the UPDATE clause, you would be updating the
> materialized table, which would then disappear and leave the base
> table untouched.
>
> The SET clause is a list of columns to be changed or made; the WHERE
> clause tells the statement which rows to use. For this discussion, we
> will assume the user doing the update has applicable UPDATE privileges
> for each <object column>.
>
> * The WHERE Clause
>
> As mentioned, the most important thing to remember about the WHERE
> clause is that it is optional. If there is no WHERE clause, all rows
> in the table are changed. This is a common error; if you make it,
> immediately execute a ROLLBACK statement.
>
> All rows that test TRUE for the <search condition> are marked as a
> subset and not as individual rows. It is also possible that this
> subset will be empty. This subset is used to construct a new set of
> rows that will be inserted into the table when the subset is deleted
> from the table. Note that the empty subset is a valid update that
> will fire declarative referential actions and triggers.
>
> * The SET Clause
>
> Each assignment in the <set clause list> is executed in parallel and
> each SET clause changes all the qualified rows at once. Or at least
> that is the theoretical model. In practice, implementations will
> first mark all of the qualified rows in the table in one pass, using
> the WHERE clause. If there were no problems, then the SQL engine
> makes a copy of each marked row in working storage. Each SET clause
> is executed based on the old row image and the results are put in the
> new row image. Finally, the old rows are deleted and the new rows are
> inserted. If an error occurs during all of this, then system does a
> ROLLBACK, the table is left unchanged and the errors are reported.
> This parallelism is not like what you find in a traditional
> third-generation programming language, so it may be hard to learn.
> This feature lets you write a statement that will swap the values in
> two columns, thus:
>
> UPDATE MyTable
> SET a = b, b = a;
>
> This is not the same thing as
>
> BEGIN ATOMIC
> UPDATE MyTable
> SET a = b;
> UPDATE MyTable
> SET b = a;
> END;
>
> In the first UPDATE, columns a and b will swap values in each row. In
> the second pair of UPDATEs, column a will get all of the values of
> column b in each row. In the second UPDATE of the pair, a, which now
> has the same value as the original value of b, will be written back
> into column b -- no change at all. There are some limits as to what
> the value expression can be. The same column cannot appear more than
> once in a <set clause list> -- which makes sense, given the parallel
> nature of the statement. Since both go into effect at the same time,
> you would not know which SET clause to use.
>
> If a subquery expression is used in a <set clause>, and it returns a
> single value, the result set is cast to a scalar; if it returns an
> empty, the result set is cast to a NULL; if it returns multiple rows,
> a cardinality violation is raised.
>
> When we finally get T-SQL up to SQL-92 specs, you could use row
> constructors and have written:
>
> UPDATE Test
> SET ROW(a,b)
> = (SELECT AVG(q), SUM(q) FROM FooWHERE Foo.p = Test.y);[/color]

[Steven An]

Re: UPDATEs with multiple aggregate functions

OK I didn't read the "When we finally" part. So as of now, my best
actual solution seems to be the JOIN'd UPDATE.. I guess I'll have to
break standard. Oh well..

jcelko212@earth link.net (--CELKO--) wrote in message news:<18c7b3c2. 0406121447.35ad ed55@posting.go ogle.com>...[color=blue][color=green][color=darkred]
> >> I need to write an update query with multiple aggregate functions.[/color][/color]
> Here is an example: <<
>
> You might want to learn Standard SQL. There is no FROM clause in a
> Standard SQL UPDATE statement; it would make no sense. Other products
> (SQL Server, Sybase and Ingres) also use the UPDATE .. FROM syntax,
> but with different semantics. So it does not port, or even worse,
> when you do move it, it trashes your database. Other programmers
> cannot read it and maintaining it is harder. And when Microsoft
> decides to change it, you will have to do a re-write. Remember the
> deprecated "*=" versus "LEFT OUTER JOIN" conversions?
>
> The correct syntax for a searched update statement is
>
> <update statement> ::=
> UPDATE <table name>
> SET <set clause list>
> [WHERE <search condition>]
>
> <set clause list> ::=
> <set clause> [{ , <set clause> }...]
>
> <set clause> ::= <object column> = <update source>
>
> <update source> ::= <value expression> | NULL | DEFAULT
>
> <object column> ::= <column name>
>
> The UPDATE clause simply gives the name of the base table or updatable
> view to be changed.
>
> Notice that no correlation name is allowed in the UPDATE clause; this
> is to avoid some self-referencing problems that could occur. But it
> also follows the data model in Standard SQL. When you give a table
> expression a correlation name, it is to act as if a materialized table
> with that correlation name has been created in the database. That
> table then is dropped at the end of the statement. If you allowed
> correlation names in the UPDATE clause, you would be updating the
> materialized table, which would then disappear and leave the base
> table untouched.
>
> The SET clause is a list of columns to be changed or made; the WHERE
> clause tells the statement which rows to use. For this discussion, we
> will assume the user doing the update has applicable UPDATE privileges
> for each <object column>.
>
> * The WHERE Clause
>
> As mentioned, the most important thing to remember about the WHERE
> clause is that it is optional. If there is no WHERE clause, all rows
> in the table are changed. This is a common error; if you make it,
> immediately execute a ROLLBACK statement.
>
> All rows that test TRUE for the <search condition> are marked as a
> subset and not as individual rows. It is also possible that this
> subset will be empty. This subset is used to construct a new set of
> rows that will be inserted into the table when the subset is deleted
> from the table. Note that the empty subset is a valid update that
> will fire declarative referential actions and triggers.
>
> * The SET Clause
>
> Each assignment in the <set clause list> is executed in parallel and
> each SET clause changes all the qualified rows at once. Or at least
> that is the theoretical model. In practice, implementations will
> first mark all of the qualified rows in the table in one pass, using
> the WHERE clause. If there were no problems, then the SQL engine
> makes a copy of each marked row in working storage. Each SET clause
> is executed based on the old row image and the results are put in the
> new row image. Finally, the old rows are deleted and the new rows are
> inserted. If an error occurs during all of this, then system does a
> ROLLBACK, the table is left unchanged and the errors are reported.
> This parallelism is not like what you find in a traditional
> third-generation programming language, so it may be hard to learn.
> This feature lets you write a statement that will swap the values in
> two columns, thus:
>
> UPDATE MyTable
> SET a = b, b = a;
>
> This is not the same thing as
>
> BEGIN ATOMIC
> UPDATE MyTable
> SET a = b;
> UPDATE MyTable
> SET b = a;
> END;
>
> In the first UPDATE, columns a and b will swap values in each row. In
> the second pair of UPDATEs, column a will get all of the values of
> column b in each row. In the second UPDATE of the pair, a, which now
> has the same value as the original value of b, will be written back
> into column b -- no change at all. There are some limits as to what
> the value expression can be. The same column cannot appear more than
> once in a <set clause list> -- which makes sense, given the parallel
> nature of the statement. Since both go into effect at the same time,
> you would not know which SET clause to use.
>
> If a subquery expression is used in a <set clause>, and it returns a
> single value, the result set is cast to a scalar; if it returns an
> empty, the result set is cast to a NULL; if it returns multiple rows,
> a cardinality violation is raised.
>
> When we finally get T-SQL up to SQL-92 specs, you could use row
> constructors and have written:
>
> UPDATE Test
> SET ROW(a,b)
> = (SELECT AVG(q), SUM(q) FROM FooWHERE Foo.p = Test.y);[/color]

[Erland Sommarskog]

Re: UPDATEs with multiple aggregate functions

Steven An (stevenan@uclin k.berkeley.edu) writes:[color=blue]
> Well, thank you for that very insightful answer. Yes, that ROW
> operator is a lot like what I had in mind. I was looking for
> something like SET (a,b) = (select ..). I will try that, and see
> which one is faster. I have a feeling your method will be MUCH
> faster, since unlike the JOIN, it does not require that I calculate
> avg's and sum's for ALL rows, only the ones I want. Thank you![/color]

Actually, my experience is that on SQL Server

UPDATE tbl
SET col = (SELECT SUM(x)
FROM tbl2
WHERE tbl2.keycol1 = tbl.keycol)

usually performs less well than:

UPDATE tbl
SET col = x.sum2
FROM tbl a
JOIN (SELECT keycol1, sum2 = SUM(x) FROM tbl2 GROUP BY keycol1) b
ON a.keycol = b.keycol1

The derived table, is only logical and SQL Server does not have to
compute all sums, even if the looks so. We had one case in our
system where a colleague rewrote from the first syntax to second,
and execution time fell from two minutes to a few seconds. But

If you have more than one column to update, using a derived table is
almost a guaranteed winner.

Of course, this is all up to the optimizer, and if you need to know
for the specific case, you have to benchmark.

--
Erland Sommarskog, SQL Server MVP, esquel@sommarsk og.se

Books Online for SQL Server SP3 at

http://www.microsoft.com/sql/techinfo/productdoc/2000/books.asp