I have some entries in my database, in my case Videos with a rating and popularity and other factors. Of all these factors I calculate a likelihood factor or more to say a boost factor.
So I essentially have the fields ID and BOOST.The boost is calculated in a way that it turns out as an integer that represents the percentage of how often this entry should be hit in in comparison.
ID Boost
1 1
2 2
3 7
So if I run my random function indefinitely I should end up with X hits on ID 1, twice as much on ID 2 and 7 times as much on ID 3.
So every hit should be random but with a probability of (boost / sum of boosts). So the probability for ID 3 in this example should be 0.7 (because the sum is 10. I choose those values for simplicity).
I thought about something like the following query:
SELECT id FROM table WHERE CEIL(RAND() * MAX(boost)) >= boost ORDER BY rand();
Unfortunately that doesn't work, after considering the following entries in the table:
ID Boost
1 1
2 2
It will, with a 50/50 chance, have only the 2nd or both elements to choose from randomly.
So 0.5 hit goes to the second element
And 0.5 hit goes to the (second and first) element which is chosen from randomly so so 0.25 each.
So we end up with a 0.25/0.75 ratio, but it should be 0.33/0.66
I need some modification or new a method to do this with good performance.
I also thought about storing the boost field cumulatively so I just do a range query from (0-sum()), but then I would have to re-index everything coming after one item if I change it or develop some swapping algorithm or something... but that's really not elegant and stuff.
Both inserting/updating and selecting should be fast!
Do you have any solutions to this problem?
The best use case to think of is probably advertisement delivery. "Please choose a random ad with given probability"... however i need it for another purpose but just to give you a last picture what it should do.
edit:
Thanks to kens answer i thought about the following approach:
calculate a random value from 0-sum(distinct boost)
SET #randval = (select ceil(rand() * sum(DISTINCT boost)) from test);
select the boost factor from all distinct boost factors which added up surpasses the random value
then we have in our 1st example 1 with a 0.1, 2 with a 0.2 and 7 with a 0.7 probability.
now select one random entry from all entries having this boost factor
PROBLEM: because the count of entries having one boost is always different. For example if there is only 1-boosted entry i get it in 1 of 10 calls, but if there are 1 million with 7, each of them is hardly ever returned...
so this doesnt work out :( trying to refine it.
I have to somehow include the count of entries with this boost factor ... but i am somehow stuck on that...
You need to generate a random number per row and weight it.
In this case, RAND(CHECKSUM(NEWID())) gets around the "per query" evaluation of RAND. Then simply multiply it by boost and ORDER BY the result DESC. The SUM..OVER gives you the total boost
DECLARE #sample TABLE (id int, boost int)
INSERT #sample VALUES (1, 1), (2, 2), (3, 7)
SELECT
RAND(CHECKSUM(NEWID())) * boost AS weighted,
SUM(boost) OVER () AS boostcount,
id
FROM
#sample
GROUP BY
id, boost
ORDER BY
weighted DESC
If you have wildly different boost values (which I think you mentioned), I'd also consider using LOG (which is base e) to smooth the distribution.
Finally, ORDER BY NEWID() is a randomness that would take no account of boost. It's useful to seed RAND but not by itself.
This sample was put together on SQL Server 2008, BTW
I dare to suggest straightforward solution with two queries, using cumulative boost calculation.
First, select sum of boosts, and generate some number between 0 and boost sum:
select ceil(rand() * sum(boost)) from table;
This value should be stored as a variable, let's call it {random_number}
Then, select table rows, calculating cumulative sum of boosts, and find the first row, which has cumulative boost greater than {random number}:
SET #cumulative_boost=0;
SELECT
id,
#cumulative_boost:=(#cumulative_boost + boost) AS cumulative_boost,
FROM
table
WHERE
cumulative_boost >= {random_number}
ORDER BY id
LIMIT 1;
My problem was similar: Every person had a calculated number of tickets in the final draw. If you had more tickets then you would have an higher chance to win "the lottery".
Since I didn't trust any of the found results rand() * multiplier or the one with -log(rand()) on the web I wanted to implement my own straightforward solution.
What I did and in your case would look a little bit like this:
(SELECT id, boost FROM foo) AS values
INNER JOIN (
SELECT id % 100 + 1 AS counter
FROM user
GROUP BY counter) AS numbers ON numbers.counter <= values.boost
ORDER BY RAND()
Since I don't have to run it often I don't really care about future performance and at the moment it was fast for me.
Before I used this query I checked two things:
The maximum number of boost is less than the maximum returned in the number query
That the inner query returns ALL numbers between 1..100. It might not depending on your table!
Since I have all distinct numbers between 1..100 then joining on numbers.counter <= values.boost would mean that if a row has a boost of 2 it would end up duplicated in the final result. If a row has a boost of 100 it would end up in the final set 100 times. Or in another words. If sum of boosts is 4212 which it was in my case you would have 4212 rows in the final set.
Finally I let MySql sort it randomly.
Edit: For the inner query to work properly make sure to use a large table, or make sure that the id's don't skip any numbers. Better yet and probably a bit faster you might even create a temporary table which would simply have all numbers between 1..n. Then you could simply use INNER JOIN numbers ON numbers.id <= values.boost
Related
I have the following MYSQL table:
measuredata:
- ID (bigint)
- timestamp
- entityid
- value (double)
The table contains >1 billion entries. I want to be able to visualize any time-window. The time window can be size of "one day" to "many years". There are measurement values round about every minute in DB.
So the number of entries for a time-window can be quite different. Say from few hundrets to several thousands or millions.
Those values are ment to be visualiuzed in a graphical chart-diagram on a webpage.
If the chart is - lets say - 800px wide, it does not make sense to get thousands of rows from database if time-window is quite big. I cannot show more than 800 values on this chart anyhow.
So, is there a way to reduce the resultset directly on DB-side?
I know "average" and "sum" etc. as aggregate function. But how can I i.e. aggregate 100k rows from a big time-window to lets say 800 final rows?
Just getting those 100k rows and let the chart do the magic is not the preferred option. Transfer-size is one reason why this is not an option.
Isn't there something on DB side I can use?
Something like avg() to shrink X rows to Y averaged rows?
Or a simple magic to just skip every #th row to shrink X to Y?
update:
Although I'm using MySQL right now, I'm not tied to this. If PostgreSQL f.i. provides a feature that could solve the issue, I'm willing to switch DB.
update2:
I maybe found a possible solution: https://mike.depalatis.net/blog/postgres-time-series-database.html
See section "Data aggregation".
The key is not to use a unixtimestamp but a date and "trunc" it, avergage the values and group by the trunc'ed date. Could work for me, but would require a rework of my table structure. Hmm... maybe there's more ... still researching ...
update3:
Inspired by update 2, I came up with this query:
SELECT (`timestamp` - (`timestamp` % 86400)) as aggtimestamp, `entity`, `value` FROM `measuredata` WHERE `entity` = 38 AND timestamp > UNIX_TIMESTAMP('2019-01-25') group by aggtimestamp
Works, but my DB/index/structue seems not really optimized for this: Query for last year took ~75sec (slow test machine) but finally got only a one value per day. This can be combined with avg(value), but this further increases query time... (~82sec). I will see if it's possible to further optimize this. But I now have an idea how "downsampling" data works, especially with aggregation in combination with "group by".
There is probably no efficient way to do this. But, if you want, you can break the rows into equal sized groups and then fetch, say, the first row from each group. Here is one method:
select md.*
from (select md.*,
row_number() over (partition by tile order by timestamp) as seqnum
from (select md.*, ntile(800) over (order by timestamp) as tile
from measuredata md
where . . . -- your filtering conditions here
) md
) md
where seqnum = 1;
I'm trying to select a random row from a table, but there is a column in this table called Rate, I want it to return the row that has a higher rate, and rarely ever return the rows that has a lower rate, is this possible?
Table :
CREATE TABLE _Random (Code varchar(128), Rate tinyint)
So you want a random row, but weighted towards the ones with higher rates?
It would also be good to know how many rows there are in the table - sorting the whole lot is kinda expensive. You may prefer to use a row_number concept than sorting by N guids.
So... One option could be to generate a single number, and then divide 100 by it. Imagine we generate a number between 0 and 1.
.25 gives us 400, .5 gives us 200, .75 gives us 133... Notice that there's a curve here - so the numbers closer to 100 come up more often (subtract 100 to make the range start at 1).
You could use RAND() for a single value between 0 and 1 (it's probably good enough), and then do the division and subtraction to get a number. If this is higher than the count of records, then maybe repeat? But try to choose a value for your division that suits.
If you need to weight it more, you could raise your RAND() value by some number, to flatten it out or steepen it up. Do some experimenting to see how it looks.
This query will fetch a random record which has an above average rate
SELECT TOP (1) * FROM _Random
WHERE Rate>(SELECT AVG(Rate) FROM _Random)
ORDER BY NEWID()
In my postgres database, I have the following relationships (simplified for the sake of this question):
Objects (currently has about 250,000 records)
-------
n_id
n_store_object_id (references store.n_id, 1-to-1 relationship, some objects don't have store records)
n_media_id (references media.n_id, 1-to-1 relationship, some objects don't have media records)
Store (currently has about 100,000 records)
-----
n_id
t_name,
t_description,
n_status,
t_tag
Media
-----
n_id
t_media_path
So far, so good. When I need to query the data, I run this (note the limit 2 at the end, as part of the requirement):
select
o.n_id,
s.t_name,
s.t_description,
me.t_media_path
from
objects o
join store s on (o.n_store_object_id = s.n_id and s.n_status > 0 and s.t_tag is not null)
join media me on o.n_media_id = me.n_id
limit
2
This works fine and gives me two entries back, as expected. The execution time on this is about 20 ms - just fine.
Now I need to get 2 random entries every time the query runs. I thought I'd add order by random(), like so:
select
o.n_id,
s.t_name,
s.t_description,
me.t_media_path
from
objects o
join store s on (o.n_store_object_id = s.n_id and s.n_status > 0 and s.t_tag is not null)
join media me on o.n_media_id = me.n_id
order by
random()
limit
2
While this gives the right results, the execution time is now about 2,500 ms (over 2 seconds). This is clearly not acceptable, as it's one of a number of queries to be run to get data for a page in a web app.
So, the question is: how can I get random entries, as above, but still keep the execution time within some reasonable amount of time (i.e. under 100 ms is acceptable for my purpose)?
Of course it needs to sort the whole thing according to random criteria before getting first rows. Maybe you can work around by using random() in offset instead?
Here's some previous work done on the topic which may prove helpful:
http://blog.rhodiumtoad.org.uk/2009/03/08/selecting-random-rows-from-a-table/
I'm thinking you'll be better off selecting random objects first, then performing the join to those objects after they're selected. I.e., query once to select random objects, then query again to join just those objects that were selected.
It seems like your problem is this: You have a table with 250,000 rows and need two random rows. Thus, you have to generate 250,000 random numbers and then sort the rows by their numbers. Two seconds to do this seems pretty fast to me.
The only real way to speed up the selection is not have to come up with 250,000 random numbers, but instead lookup rows through an index.
I think you'd have to change the table schema to optimize for this case. How about something like:
1) Create a new column with a sequence starting at 1.
2) Every row will then have a number.
3) Create an index on: number % 1000
4) Query for rows where number % 1000 is equal to a random number
between 0 and 999 (this should hit the index and load a random
portion of your database)
5) You can probably then add on RANDOM() to your ORDER BY clause and
it will then just sort that chunk of your database and be 1,000x
faster.
6) Then select the first two of those rows.
If this still isn't random enough (since rows will always be paired having the same "hash"), you could probably do a union of two random rows, or have an OR clause in the query and generate two random keys.
Hopefully something along these lines could be very fast and decently random.
I have a table foodbar, created with the following DDL. (I am using mySQL 5.1.x)
CREATE TABLE foodbar (
id INT NOT NULL AUTO_INCREMENT,
user_id INT NOT NULL,
weight double not null,
created_at date not null
);
I have four questions:
How may I write a query that returns
a result set that gives me the
following information: user_id,
weight_gain where weight_gain is
the difference between a weight and
a weight that was recorded 7 days
ago.
How may I write a query that will
return the top N users with the
biggest weight gain (again say over
a week).? An 'obvious' way may be to
use the query obtained in question 1
above as a subquery, but somehow
picking the top N.
Since in question 2 (and indeed
question 1), I am searching the
records in the table using a
calculated field, indexing would be
preferable to optimise the query -
however since it is a calculated
field, it is not clear which field
to index (I'm guessing the 'weight'
field is the one that needs
indexing). Am I right in that
assumption?.
Assuming I had another field in the
foodbar table (say 'height') and I
wanted to select records from the
table based on (say) the product
(i.e. multiplication) of 'height'
and 'weight' - would I be right in
assuming again that I need to index
'height' and 'weight'?. Do I also
need to create a composite key (say
(height,weight)). If this question
is not clear, I would be happy to
clarify
I don't see why you should need the synthetic key, so I'll use this table instead:
CREATE TABLE foodbar (
user_id INT NOT NULL
, created_at date not null
, weight double not null
, PRIMARY KEY (user_id, created_at)
);
How may I write a query that returns a result set that gives me the following information: user_id, weight_gain where weight_gain is the difference between a weight and a weight that was recorded 7 days ago.
SELECT curr.user_id, curr.weight - prev.weight
FROM foodbar curr, foodbar prev
WHERE curr.user_id = prev.user_id
AND curr.created_at = CURRENT_DATE
AND prev.created_at = CURRENT_DATE - INTERVAL '7 days'
;
the date arithmetic syntax is probably wrong but you get the idea
How may I write a query that will return the top N users with the biggest weight gain (again say over a week).? An 'obvious' way may be to use the query obtained in question 1 above as a subquery, but somehow picking the top N.
see above, add ORDER BY curr.weight - prev.weight DESC and LIMIT N
for the last two questions: don't speculate, examine execution plans. (postgresql has EXPLAIN ANALYZE, dunno about mysql) you'll probably find you need to index columns that participate in WHERE and JOIN, not the ones that form the result set.
I think that "just somebody" covered most of what you're asking, but I'll just add that indexing columns that take part in a calculation is unlikely to help you at all unless it happens to be a covering index.
For example, it doesn't help to order the following rows by X, Y if I want to get them in the order of their product X * Y:
X Y
1 8
2 2
4 4
The products would order them as:
X Y Product
2 2 4
1 8 8
4 4 16
If mySQL supports calculated columns in a table and allows indexing on those columns then that might help.
I agree with just somebody regarding the primary key, but for what you're asking regarding the weight calculation, you'd be better off storing the delta rather than the weight:
CREATE TABLE foodbar (
user_id INT NOT NULL,
created_at date not null,
weight_delta double not null,
PRIMARY KEY (user_id, created_at)
);
It means you'd store the users initial weight in say, the user table, and when you write records to the foodbar table, a user could supply the weight at that time, but the query would subtract the initial weight from the current weight. So you'd see values like:
user_id weight_delta
------------------------
1 2
1 5
1 -3
Looking at that, you know that user 1 gained 4 pounds/kilos/stones/etc.
This way you could use SUM, because it's possible for someone to have weighings every day - using just somebody's equation of curr.weight - prev.weight wouldn't work, regardless of time span.
Getting the top x is easy in MySQL - use the LIMIT clause, but mind that you provide an ORDER BY to make sure the limit is applied correctly.
It's not obvious, but there's some important information missing in the problem you're trying to solve. It becomes more noticeable when you think about realistic data going into this table. The problem is that you're unlikely to to have a consistent regular daily record of users' weights. So you need to clarify a couple of rules around determining 'current-weight' and 'weight x days ago'. I'm going to assume the following simplistic rules:
The most recent weight reading is the 'current-weight'. (Even though that could be months ago.)
The most recent weight reading more than x days ago will be the weight assumed at x days ago. (Even though for example a reading from 6 days ago would be more reliable than a reading from 21 days ago when determining weight 7 days ago.)
Now to answer the questions:
1&2: Using the above extra rules provides an opportunity to produce two result sets: current weights, and previous weights:
Current weights:
select rd.*,
w.Weight
from (
select User_id,
max(Created_at) AS Read_date
from Foodbar
group by User_id
) rd
inner join Foodbar w on
w.User_id = rd.User_id
and w.Created_at = rd.Read_date
Similarly for the x days ago reading:
select rd.*,
w.Weight
from (
select User_id,
max(Created_at) AS Read_date
from Foodbar
where Created_at < DATEADD(dd, -7, GETDATE()) /*Or appropriate MySql equivalent*/
group by User_id
) rd
inner join Foodbar w on
w.User_id = rd.User_id
and w.Created_at = rd.Read_date
Now simply join these results as subqueries
select cur.User_id,
cur.Weight as Cur_weight,
prev.Weight as Prev_weight
cur.Weight - prev.Weight as Weight_change
from (
/*Insert query #1 here*/
) cur
inner join (
/*Insert query #2 here*/
) prev on
prev.User_id = cur.User_id
If I remember correctly the MySql syntax to get the top N weight gains would be to simply add:
ORDER BY cur.Weight - prev.Weight DESC limit N
2&3: Choosing indexes requires a little understanding of how the query optimiser will process the query:
The important thing when it comes to index selection is what columns you are filtering by or joining on. The optimiser will use the index if it is determined to be selective enough (note that sometimes your filters have to be extremely selective returning < 1% of data to be considered useful). There's always a trade of between slow disk seek times of navigating indexes and simply processing all the data in memory.
3: Although weights feature significantly in what you display, the only relevance is in terms of filtering (or selection) is in #2 to get the top N weight gains. This is a complex calculation based on a number of queries and a lot of processing that has gone before; so Weight will provide zero benefit as an index.
Another note is that even for #2 you have to calculate the weight change of all users in order to determine the which have gained the most. Therefore unless you have a very large number of readings per user you will read most of the table. (I.e. a table scan will be used to obtain the bulk of the data)
Where indexes can benefit:
You are trying to identify specific Foodbar rows based on User_id and Created_at.
You are also joining back to the Foodbar table again using User_id and Created_at.
This implies an index on User_id, Created__at would be useful (more-so if this is the clustered index).
4: No, unfortunately it is mathematically impossible to determine how the individual values H and W would independently determine the ordering of the product. E.g. both H=3 & W=3 are less than 5, yet if H=5 and W=1 then the product 3*3 is greater than 5*1.
You would have to actually store the calculation an index on that additional column. However, as indicated in my answer to #3 above, it is still unlikely to prove beneficial.
I want to create a table, with each row containing some sort of weight. Then I want to select random values with the probability equal to (weight of that row)/(weight of all rows). For example, having 5 rows with weights 1,2,3,4,5 out of 1000 I'd get approximately 1/15*1000=67 times first row and so on.
The table is to be filled manually. Then I'll take a random value from it. But I want to have an ability to change the probabilities on the filling stage.
I found this nice little algorithm in Quod Libet. You could probably translate it to some procedural SQL.
function WeightedShuffle(list of items with weights):
max_score ← the sum of every item’s weight
choice ← random number in the range [0, max_score)
current ← 0
for each item (i, weight) in items:
current ← current + weight
if current ≥ choice or i is the last item:
return item i
The easiest (and maybe best/safest?) way to do this is to add those rows to the table as many times as you want the weight to be - say I want "Tree" to be found 2x more often then "Dog" - I insert it 2 times into the table and I insert "Dog" once and just select elements at random one by one.
If the rows are complex/big then it would be best to create a separate table (weighted_Elements or something) in which you'll just have foreign keys to the real rows inserted as many times as the weights dictate.
The best possible scenario (if i understand your question properly) is to setup your table as you normally would and then add two columns both INT's.
Column 1: Weight - This column would hold your weight value going from -X to +X, X being the highest value you want to have as a weight (IE: X=100, -100 to 100). This value is populated to give the row an actual weight and increase or decrease the probability of it coming up.
Column 2: *Count** - This column would hold the count of how many times this row has come up, this column is needed only if you want to use fair weighting. Fair weighting prevents one row from always showing up. (IE: if you have one row weighted at 100 and another at 2 the row with 100 will always show up, this column will allow weight 2 to be more 'valueable' as you get more weight 100 results). This column should be incremented by 1 each time a row result is pulled but you can make the logic more advanced later so it adds the weight etc.
Logic: - Its really simple now, your query simply has to request all rows as you normally would then make an extra select that (you can change the logic here to whatever you want) takes the weights and subtracts the count and order by that column.
The end result should be a table where you will get your weights appearing more often until a certain point where the system will evenly distribute itself out (leave out column 2) and you will have a system that will always return the same weighted order unless you offset the base of the query (IE: LIMIT [RANDOM NUMBER], [NUMBER OF ROWS TO RETURN])
I'm not an expert in probability theory, but assuming you have a column called WEIGHT, how about
select FIELD_1, ... FIELD_N, (rand() * WEIGHT) as SCORE
from YOURTABLE
order by SCORE
limit 0, 10
This would give you 10 records, but you can change the limit clause, of course.
The problem is called Reservoir Sampling (https://en.wikipedia.org/wiki/Reservoir_sampling)
The A-Res algorithm is easy to implement in SQL:
SELECT *
FROM table
ORDER BY pow(rand(), 1 / weight) DESC
LIMIT 10;
I came looking for the answer to the same question - I decided to come up with this:
id weight
1 5
2 1
SELECT * FROM table ORDER BY RAND()/weight
it's not exact - but it is using random so i might not expect exact. I ran it 70 times to get number 2 10 times. I would have expect 1/6th but i got 1/7th. I'd say that's pretty close. I'd have to run a script to do it a few thousand times to get a really good idea if it's working.