proposal: window function - change_number

On Sun, Sep 21, 2014 at 9:27 PM, Pavel Stehule <pavel.stehule@gmail.com>
wrote:

Hi
I tried to solve following task:

I have a table

start, reason, km
=============
2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20
2014-01-01 10:00:00, commerc, 20
2014-01-01 11:00:00, private, 8

and I would reduce these rows to

2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20 + 20 = 40
2014-01-01 11:00:00, private, 8

It is relative hard to it now with SQL only. But we can simplify this task
with window function that returns number of change in some column. Then
this task can be solved by

select min(start), min(reason), sum(km)
from (select start, reason, km, change_number(reason) over (order by
start))
group by change_number;

I guess that might be quite useful, otherwise the only way that comes to
mind to do this would be something along the lines of:

select *,sum(case when reason <> lastreason then 1 else 0 end) over (order
by start) as chg_num from (select *,lag(reason) over (order by start) vnext
from sometable) sometable;

This way might not be too bad as I think the outer window will have no need
to perform another sort, since the inner window clause has sorted it the
right way already. Though something like change_number() would make this a
bit more pretty. It's almost like rank(), but with a parameter.

Regards

David Rowley

Hi Pavel (and others),

Pavel Stehule wrote:

Hi
I tried to solve following task:

I have a table

start, reason, km
=============
2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20
2014-01-01 10:00:00, commerc, 20
2014-01-01 11:00:00, private, 8

and I would reduce these rows to

2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20 + 20 = 40
2014-01-01 11:00:00, private, 8

It is relative hard to it now with SQL only. But we can simplify this task
with window function that returns number of change in some column. Then
this task can be solved by

select min(start), min(reason), sum(km)
from (select start, reason, km, change_number(reason) over (order by
start))
group by change_number;

What about

select srk.reason, min(srk.start), sum(srk.km)
from start_reason_km srk
group by srk.reason, (select max(start) from start_reason_km other WHERE
other.start < srk.start and other.reason != srk.reason);

In general, I think window function are very specific in how the queryplan
must look like, leaving not much room for the optimizer. On the other hand,
if there happends to be an efficient way to get the results of the table
ordered by "start", then the window function will very likely much faster
then a join. I would be nice if the optimizer is able to add such stream
order operations.

Do you think, so it has sense?

Regards

Pavel

Regards,

Mart

PS: This is my first post to the mailing list. I am a software developer
interest is performance making webapplications with a different database
server during working hours.

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

2014-09-21 14:30 GMT+02:00 Mart Kelder <mart@kelder31.nl>:

Hi Pavel (and others),

Pavel Stehule wrote:

Hi
I tried to solve following task:

I have a table

start, reason, km
=============
2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20
2014-01-01 10:00:00, commerc, 20
2014-01-01 11:00:00, private, 8

and I would reduce these rows to

2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20 + 20 = 40
2014-01-01 11:00:00, private, 8

It is relative hard to it now with SQL only. But we can simplify this

task

with window function that returns number of change in some column. Then
this task can be solved by

select min(start), min(reason), sum(km)
from (select start, reason, km, change_number(reason) over (order by
start))
group by change_number;

What about

select srk.reason, min(srk.start), sum(srk.km)
from start_reason_km srk
group by srk.reason, (select max(start) from start_reason_km other WHERE
other.start < srk.start and other.reason != srk.reason);

This query is Cartesian product, so for some large data it is significantly
slower then window function (required only sorts without joins)

My motivation was a) to implement described task without Cartesian product.
b) introduce some tool for this kind of problems. I seen more times a
request .. reduce a time series, and a window function "change_number" (or
maybe "consistent_series_number") can be good candidate.

Hi Pavel (and others),

Op zondag 21 september 2014 15:35:46 schreef u:

2014-09-21 14:30 GMT+02:00 Mart Kelder <mart@kelder31.nl>:

Hi Pavel (and others),

Pavel Stehule wrote:

Hi
I tried to solve following task:

I have a table

start, reason, km
=============

2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20
2014-01-01 10:00:00, commerc, 20
2014-01-01 11:00:00, private, 8

and I would reduce these rows to

2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20 + 20 = 40
2014-01-01 11:00:00, private, 8

It is relative hard to it now with SQL only. But we can simplify this
task

with window function that returns number of change in some column. Then
this task can be solved by

select min(start), min(reason), sum(km)
from (select start, reason, km, change_number(reason) over (order by
start))

group by change_number;

What about

select srk.reason, min(srk.start), sum(srk.km)
from start_reason_km srk
group by srk.reason, (select max(start) from start_reason_km other WHERE
other.start < srk.start and other.reason != srk.reason);

This query is Cartesian product, so for some large data it is significantly
slower then window function (required only sorts without joins)

I think we have the same queryplan in mind (with only one scan). As far as I
know, SQL is a language where you define the result you want to get, and let
the server find a way how to find the data. I think windowed function also say
how the server needs to get the information.

The real challenge is of course of finding heuristics to remove the additional
join. In this particular case, I can tell how to remove the inner join from
the subquery:
* the where-clause of the self-join contains other.start < srk.start. From
that we can conclude that if the table is (or can be) sorted on "start", we
have seen the data before the subquery is executed
* because we only need an aggregate, we need to store the intermediate "max"
for each "reason". And then add the result to the stream.

Recently, I had a simular problem with a table containing a timestamp, a state
and a object where the state belongs to. A object remains in a state until
there is a more recent tuple in the table. I needed basically to query all the
previous state for each tuple, but preverably without the additional join.

My motivation was a) to implement described task without Cartesian product.

Good reason (if you consider the queryplan and not the query).

b) introduce some tool for this kind of problems. I seen more times a
request .. reduce a time series, and a window function "change_number" (or
maybe "consistent_series_number") can be good candidate.

I also need to note that there is a lot of difference in complexity between the
possible solutions of this problem. Where a new window function can probably
be very easy implemented, the optimizer changes descripted above are complex
and not easy to implement.

I also want to note that I am not really against new window functions, I only
want to point out that a more generic solution also might be possible.

Regards,

Mart

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

2014-09-21 17:00 GMT+02:00 Mart Kelder <mart@kelder31.nl>:

Hi Pavel (and others),

Op zondag 21 september 2014 15:35:46 schreef u:

2014-09-21 14:30 GMT+02:00 Mart Kelder <mart@kelder31.nl>:

Hi Pavel (and others),

Pavel Stehule wrote:

Hi
I tried to solve following task:

I have a table

start, reason, km
=============

2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20
2014-01-01 10:00:00, commerc, 20
2014-01-01 11:00:00, private, 8

and I would reduce these rows to

2014-01-01 08:00:00, private, 10
2014-01-01 09:00:00, commerc, 20 + 20 = 40
2014-01-01 11:00:00, private, 8

It is relative hard to it now with SQL only. But we can simplify this
task

with window function that returns number of change in some column.

Then

this task can be solved by

select min(start), min(reason), sum(km)
from (select start, reason, km, change_number(reason) over (order

by

start))

group by change_number;

What about

select srk.reason, min(srk.start), sum(srk.km)
from start_reason_km srk
group by srk.reason, (select max(start) from start_reason_km other

WHERE

other.start < srk.start and other.reason != srk.reason);

This query is Cartesian product, so for some large data it is

significantly

slower then window function (required only sorts without joins)

I think we have the same queryplan in mind (with only one scan). As far as
I
know, SQL is a language where you define the result you want to get, and
let
the server find a way how to find the data. I think windowed function also
say
how the server needs to get the information.

What I know It is not true now. Any subselect enforce individual scan of
source relation. Postgres has no any special support for selfjoin.

The real challenge is of course of finding heuristics to remove the
additional
join. In this particular case, I can tell how to remove the inner join from
the subquery:
* the where-clause of the self-join contains other.start < srk.start. From
that we can conclude that if the table is (or can be) sorted on "start", we
have seen the data before the subquery is executed
* because we only need an aggregate, we need to store the intermediate
"max"
for each "reason". And then add the result to the stream.

Recently, I had a simular problem with a table containing a timestamp, a
state
and a object where the state belongs to. A object remains in a state until
there is a more recent tuple in the table. I needed basically to query all
the
previous state for each tuple, but preverably without the additional join.

My motivation was a) to implement described task without Cartesian

product.

Good reason (if you consider the queryplan and not the query).

yes.

There is probably big space for improvements in more directions. For
example - we have application, where is often used pattern SELECT FROM A
JOIN (SELECT someagg() FROM A) .. ON

Sometimes these queries are slow due terrible low estimation. It is one
example of more

Pavel

"Pavel" == Pavel Stehule <pavel.stehule@gmail.com> writes:

Pavel> Hi
Pavel> I tried to solve following task:

Pavel> I have a table

Pavel> start, reason, km
Pavel> =============
Pavel> 2014-01-01 08:00:00, private, 10
Pavel> 2014-01-01 09:00:00, commerc, 20
Pavel> 2014-01-01 10:00:00, commerc, 20
Pavel> 2014-01-01 11:00:00, private, 8

Pavel> and I would reduce these rows to

Pavel> 2014-01-01 08:00:00, private, 10
Pavel> 2014-01-01 09:00:00, commerc, 20 + 20 = 40
Pavel> 2014-01-01 11:00:00, private, 8

Pavel> It is relative hard to it now with SQL only.

Only relatively. My standard solution is something like this:

select start_time, reason, sum(km) as km
from (select max(label_time) over (order by start) as start_time,
reason, km
from (select start, reason, km,
case when reason
is distinct from
lag(reason) over (order by start)
then start
end as label_time
from yourtable
) s2
) s1
group by start_time, reason
order by start_time;

(Your change_number idea is essentially equivalent to doing
sum(case when x is distinct from lag(x) over w then 1 end) over w,
except that since window functions can't be nested, that expression
requires a subquery.)

--
Andrew (irc:RhodiumToad)

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Hey, sorry I what I say is obvious for you .

If I understood your problem correctly, it is strictly equivalent to this
one :
http://postgresql.1045698.n5.nabble.com/Count-of-records-in-a-row-td5775363.html

there is a postgres trick to solve this problem :
what you want is essentially generate a unique group_id,
but one that depends of an order of row not defined in the group.

The solution
is to generate a row number by the order you want , then a row number by
the group ,
then a subtraction of the 2 row number gives you an unique id per group.

The cost is that you have to use 2 windows function., hence 2 scans I guess.

Cheers,
Rémi-C

2014-09-21 17:51 GMT+02:00 Andrew Gierth <andrew@tao11.riddles.org.uk>:

2014-09-21 17:51 GMT+02:00 Andrew Gierth <andrew@tao11.riddles.org.uk>:

"Pavel" == Pavel Stehule <pavel.stehule@gmail.com> writes:

Pavel> Hi
Pavel> I tried to solve following task:

Pavel> I have a table

Pavel> start, reason, km
Pavel> =============
Pavel> 2014-01-01 08:00:00, private, 10
Pavel> 2014-01-01 09:00:00, commerc, 20
Pavel> 2014-01-01 10:00:00, commerc, 20
Pavel> 2014-01-01 11:00:00, private, 8

Pavel> and I would reduce these rows to

Pavel> 2014-01-01 08:00:00, private, 10
Pavel> 2014-01-01 09:00:00, commerc, 20 + 20 = 40
Pavel> 2014-01-01 11:00:00, private, 8

Pavel> It is relative hard to it now with SQL only.

Only relatively. My standard solution is something like this:

select start_time, reason, sum(km) as km
from (select max(label_time) over (order by start) as start_time,
reason, km
from (select start, reason, km,
case when reason
is distinct from
lag(reason) over (order by start)
then start
end as label_time
from yourtable
) s2
) s1
group by start_time, reason
order by start_time;

(Your change_number idea is essentially equivalent to doing
sum(case when x is distinct from lag(x) over w then 1 end) over w,
except that since window functions can't be nested, that expression
requires a subquery.)

yes, I found this solution in third iteration too.

so this proposal lost a main benefit

Regards

Pavel

2014-09-21 18:08 GMT+02:00 Rémi Cura <remi.cura@gmail.com>:

Hey, sorry I what I say is obvious for you .

If I understood your problem correctly, it is strictly equivalent to this
one :

http://postgresql.1045698.n5.nabble.com/Count-of-records-in-a-row-td5775363.html

there is a postgres trick to solve this problem :
what you want is essentially generate a unique group_id,
but one that depends of an order of row not defined in the group.

The solution
is to generate a row number by the order you want , then a row number by
the group ,
then a subtraction of the 2 row number gives you an unique id per group.

The cost is that you have to use 2 windows function., hence 2 scans I
guess.

yes, it is little bit similar - I found a pattern described by Andrew is
well too.

regards

Pavel