Express Nearest Neighbor Join in Postgresql? - postgresql

I have two tables Q and T, both containing a column of float numbers.
What I want to do is, for each number in Q, I want to find a number in T that has the smallest distance to it.
For example, for T={1,7,9} and Q={2,6,10}, I want to return Q,T pairs as {(2,1),(6,7),(10,9)}.
How should I express this query with SQL?
In addition, is that possible to accelerate this join by index, e.g. add an operator class which bind "FOR ORDER BY <->" with fabs calculation?

create table t (val_t integer);
create table q (val_q integer);
insert into t values (1),(7),(9);
insert into q values (2),(6),(10);
Start with a query that cross joins the two tables and adds a rank based on the difference:
SELECT val_q, val_t, rank() OVER (PARTITION BY val_q ORDER BY abs(val_t - val_q))
FROM t
JOIN q ON true ;
Use this query in a cte or subquery and filter by rank:
WITH src AS(
SELECT val_q, val_t, rank() OVER (PARTITION BY val_q ORDER BY abs(val_t - val_q))
FROM t
JOIN q ON true )
SELECT val_q, val_t FROM src
WHERE rank = 1;
val_q | val_t
-------+-------
2 | 1
6 | 7
10 | 9
See https://www.postgresql.org/docs/12/tutorial-window.html

Given this schema:
create table t (tn float);
insert into t values (1), (7), (9);
create table q (qn float);
insert into q values (2), (6), (10);
DISTINCT ON is the most straightforward way:
select distinct on (qn) qn, tn
from q
cross join t
order by qn, abs(qn - tn);
Exploiting a numeric range may perform better depending on your data sizes. If performance is an issue, then you can create an actual temp table for the range_tn CTE and put a gist index on it:
with all_tn as (
select tn
from t
union select null
), range_tn as (
select numrange(tn::numeric, (lead(tn) over w)::numeric, '[]') as tr
from all_tn
window w as (order by tn nulls first)
)
select qn,
case
when lower_inf(tr) then upper(tr)
when upper_inf(tr) then lower(tr)
when 2 * qn - lower(tr) - upper(tr) > 0 then upper(tr)
else lower(tr)
end as tn
from q
join range_tn
on qn::numeric <# tr;
Fiddle here

Related

How can I SUM distinct records in a Postgres database where there are duplicate records?

Imagine a table that looks like this:
The SQL to get this data was just SELECT *
The first column is "row_id" the second is "id" - which is the order ID and the third is "total" - which is the revenue.
I'm not sure why there are duplicate rows in the database, but when I do a SUM(total), it's including the second entry in the database, even though the order ID is the same, which is causing my numbers to be larger than if I select distinct(id), total - export to excel and then sum the values manually.
So my question is - how can I SUM on just the distinct order IDs so that I get the same revenue as if I exported to excel every distinct order ID row?
Thanks in advance!
Easy - just divide by the count:
select id, sum(total) / count(id)
from orders
group by id
See live demo.
Also handles any level of duplication, eg triplicates etc.
You can try something like this (with your example):
Table
create table test (
row_id int,
id int,
total decimal(15,2)
);
insert into test values
(6395, 1509, 112), (22986, 1509, 112),
(1393, 3284, 40.37), (24360, 3284, 40.37);
Query
with distinct_records as (
select distinct id, total from test
)
select a.id, b.actual_total, array_agg(a.row_id) as row_ids
from test a
inner join (select id, sum(total) as actual_total from distinct_records group by id) b
on a.id = b.id
group by a.id, b.actual_total
Result
| id | actual_total | row_ids |
|------|--------------|------------|
| 1509 | 112 | 6395,22986 |
| 3284 | 40.37 | 1393,24360 |
Explanation
We do not know what the reasons is for orders and totals to appear more than one time with different row_id. So using a common table expression (CTE) using the with ... phrase, we get the distinct id and total.
Under the CTE, we use this distinct data to do totaling. We join ID in the original table with the aggregation over distinct values. Then we comma-separate row_ids so that the information looks cleaner.
SQLFiddle example
http://sqlfiddle.com/#!15/72639/3
Create custom aggregate:
CREATE OR REPLACE FUNCTION sum_func (
double precision, pg_catalog.anyelement, double precision
)
RETURNS double precision AS
$body$
SELECT case when $3 is not null then COALESCE($1, 0) + $3 else $1 end
$body$
LANGUAGE 'sql';
CREATE AGGREGATE dist_sum (
pg_catalog."any",
double precision)
(
SFUNC = sum_func,
STYPE = float8
);
And then calc distinct sum like:
select dist_sum(distinct id, total)
from orders
SQLFiddle
You can use DISTINCT in your aggregate functions:
SELECT id, SUM(DISTINCT total) FROM orders GROUP BY id
Documentation here: https://www.postgresql.org/docs/9.6/static/sql-expressions.html#SYNTAX-AGGREGATES
If we can trust that the total for 1 order is actually 1 row. We could eliminate the duplicates in a sub-query by selecting the the MAX of the PK id column. An example:
CREATE TABLE test2 (id int, order_id int, total int);
insert into test2 values (1,1,50);
insert into test2 values (2,1,50);
insert into test2 values (5,1,50);
insert into test2 values (3,2,100);
insert into test2 values (4,2,100);
select order_id, sum(total)
from test2 t
join (
select max(id) as id
from test2
group by order_id) as sq
on t.id = sq.id
group by order_id
sql fiddle
In difficult cases:
select
id,
(
SELECT SUM(value::int4)
FROM jsonb_each_text(jsonb_object_agg(row_id, total))
) as total
from orders
group by id
I would suggest just use a sub-Query:
SELECT "a"."id", SUM("a"."total")
FROM (SELECT DISTINCT ON ("id") * FROM "Database"."Schema"."Table") AS "a"
GROUP BY "a"."id"
The Above will give you the total of each id
Use below if you want the full total of each duplicate removed:
SELECT SUM("a"."total")
FROM (SELECT DISTINCT ON ("id") * FROM "Database"."Schema"."Table") AS "a"
Using subselect (http://sqlfiddle.com/#!7/cef1c/51):
select sum(total) from (
select distinct id, total
from orders
)
Using CTE (http://sqlfiddle.com/#!7/cef1c/53):
with distinct_records as (
select distinct id, total from orders
)
select sum(total) from distinct_records;

Grouping consecutive dates in PostgreSQL

I have two tables which I need to combine as sometimes some dates are found in table A and not in table B and vice versa. My desired result is that for those overlaps on consecutive days be combined.
I'm using PostgreSQL.
Table A
id startdate enddate
--------------------------
101 12/28/2013 12/31/2013
Table B
id startdate enddate
--------------------------
101 12/15/2013 12/15/2013
101 12/16/2013 12/16/2013
101 12/28/2013 12/28/2013
101 12/29/2013 12/31/2013
Desired Result
id startdate enddate
-------------------------
101 12/15/2013 12/16/2013
101 12/28/2013 12/31/2013
Right. I have a query that I think works. It certainly works on the sample records you provided. It uses a recursive CTE.
First, you need to merge the two tables. Next, use a recursive CTE to get the sequences of overlapping dates. Finally, get the start and end dates, and join back to the "merged" table to get the id.
with recursive allrecords as -- this merges the input tables. Add a unique row identifier
(
select *, row_number() over (ORDER BY startdate) as rowid from
(select * from table1
UNION
select * from table2) a
),
path as ( -- the recursive CTE. This gets the sequences
select rowid as parent,rowid,startdate,enddate from allrecords a
union
select p.parent,b.rowid,b.startdate,b.enddate from allrecords b join path p on (p.enddate + interval '1 day')>=b.startdate and p.startdate <= b.startdate
)
SELECT id,g.startdate,g.enddate FROM -- outer query to get the id
-- inner query to get the start and end of each sequence
(select parent,min(startdate) as startdate, max(enddate) as enddate from
(
select *, row_number() OVER (partition by rowid order by parent,startdate) as row_number from path
) a
where row_number = 1 -- We only want the first occurrence of each record
group by parent)g
INNER JOIN allrecords a on a.rowid = parent
The below fragment does what you intend. (but it will probably be very slow) The problem is that detecteng (non)overlapping dateranges is impossible with standard range operators, since a range could be split into two parts.
So, my code does the following:
split the dateranges from table_A into atomic records, with one date per record
[the same for table_b]
cross join these two tables (we are only interested in A_not_in_B, and B_not_in_A) , remembering which of the L/R outer join wings it came from.
re-aggregate the resulting records into date ranges.
-- EXPLAIN ANALYZE
--
WITH RECURSIVE ranges AS (
-- Chop up the a-table into atomic date units
WITH ar AS (
SELECT generate_series(a.startdate,a.enddate , '1day'::interval)::date AS thedate
, 'A'::text AS which
, a.id
FROM a
)
-- Same for the b-table
, br AS (
SELECT generate_series(b.startdate,b.enddate, '1day'::interval)::date AS thedate
, 'B'::text AS which
, b.id
FROM b
)
-- combine the two sets, retaining a_not_in_b plus b_not_in_a
, moments AS (
SELECT COALESCE(ar.id,br.id) AS id
, COALESCE(ar.which, br.which) AS which
, COALESCE(ar.thedate, br.thedate) AS thedate
FROM ar
FULL JOIN br ON br.id = ar.id AND br.thedate = ar.thedate
WHERE ar.id IS NULL OR br.id IS NULL
)
-- use a recursive CTE to re-aggregate the atomic moments into ranges
SELECT m0.id, m0.which
, m0.thedate AS startdate
, m0.thedate AS enddate
FROM moments m0
WHERE NOT EXISTS ( SELECT * FROM moments nx WHERE nx.id = m0.id AND nx.which = m0.which
AND nx.thedate = m0.thedate -1
)
UNION ALL
SELECT rr.id, rr.which
, rr.startdate AS startdate
, m1.thedate AS enddate
FROM ranges rr
JOIN moments m1 ON m1.id = rr.id AND m1.which = rr.which AND m1.thedate = rr.enddate +1
)
SELECT * FROM ranges ra
WHERE NOT EXISTS (SELECT * FROM ranges nx
-- suppress partial subassemblies
WHERE nx.id = ra.id AND nx.which = ra.which
AND nx.startdate = ra.startdate
AND nx.enddate > ra.enddate
)
;

how to do dead reckoning on column of table, postgresql

I have a table looks like,
x y
1 2
2 null
3 null
1 null
11 null
I want to fill the null value by conducting a rolling
function to apply y_{i+1}=y_{i}+x_{i+1} with sql as simple as possible (inplace)
so the expected result
x y
1 2
2 4
3 7
1 8
11 19
implement in postgresql. I may encapsulate it in a window function, but the implementation of custom function seems always complex
WITH RECURSIVE t AS (
select x, y, 1 as rank from my_table where y is not null
UNION ALL
SELECT A.x, A.x+ t.y y , t.rank + 1 rank FROM t
inner join
(select row_number() over () rank, x, y from my_table ) A
on t.rank+1 = A.rank
)
SELECT x,y FROM t;
You can iterate over rows using a recursive CTE. But in order to do so, you need a way to jump from row to row. Here's an example using an ID column:
; with recursive cte as
(
select id
, y
from Table1
where id = 1
union all
select cur.id
, prev.y + cur.x
from Table1 cur
join cte prev
on cur.id = prev.id + 1
)
select *
from cte
;
You can see the query at SQL Fiddle. If you don't have an ID column, but you do have another way to order the rows, you can use row_number() to get an ID:
; with recursive sorted as
(
-- Specify your ordering here. This example sorts by the dt column.
select row_number() over (order by dt) as id
, *
from Table1
)
, cte as
(
select id
, y
from sorted
where id = 1
union all
select cur.id
, prev.y + cur.x
from sorted cur
join cte prev
on cur.id = prev.id + 1
)
select *
from cte
;
Here's the SQL Fiddle link.

Postgresql running sum of previous groups?

Given the following data:
sequence | amount
1 100000
1 20000
2 10000
2 10000
I'd like to write a sql query that gives me the sum of the current sequence, plus the sum of the previous sequence. Like so:
sequence | current | previous
1 120000 0
2 20000 120000
I know the solution likely involves windowing functions but I'm not too sure how to implement it without subqueries.
SQL Fiddle
select
seq,
amount,
lag(amount::int, 1, 0) over(order by seq) as previous
from (
select seq, sum(amount) as amount
from sa
group by seq
) s
order by seq
If your sequence is "sequencial" without holes you can simply do:
SELECT t1.sequence,
SUM(t1.amount),
(SELECT SUM(t2.amount) from mytable t2 WHERE t2.sequence = t1.sequence - 1)
FROM mytable t1
GROUP BY t1.sequence
ORDER BY t1.sequence
Otherwise, instead of t2.sequence = t1.sequence - 1 you could do:
SELECT t1.sequence,
SUM(t1.amount),
(SELECT SUM(t2.amount)
from mytable t2
WHERE t2.sequence = (SELECT MAX(t3.sequence)
FROM mytable t3
WHERE t3.sequence < t1.sequence))
FROM mytable t1
GROUP BY t1.sequence
ORDER BY t1.sequence;
You can see both approaches in this fiddle

query for a range of records in result

I am wondering if there is some easy way, a function, or other method to return data from a query with the following results.
I have a SQL Express DB 2008 R2, a table that contains numerical data in a given column, say col T.
I am given a value X in code and would like to return up to three records. The record where col T equals my value X, and the record before and after, and nothing else. The sort is done on col T. The record before may be beginning of file and therefore not exist, likewise, if X equals the last record then the record after would be non existent, end of file/table.
The value of X may not exist in the table.
This I think is similar to get a range of results in numerical order.
Any help or direction in solving this would be greatly appreciated.
Thanks again,
It might not be the most optimal solution, but:
SELECT T
FROM theTable
WHERE T = X
UNION ALL
SELECT *
FROM
(
SELECT TOP 1 T
FROM theTable
WHERE T > X
ORDER BY T
) blah
UNION ALL
SELECT *
FROM
(
SELECT TOP 1 T
FROM theTable
WHERE T < X
ORDER BY T DESC
) blah2
DECLARE #x int = 100
;WITH t as
(
select ROW_NUMBER() OVER (ORDER BY T ASC) AS row_nm,*
from YourTable
)
, t1 as
(
select *
from t
WHERE T = #x
)
select *
from t
CROSS APPLY t1
WHERE t.row_nm BETWEEN t1.row_nm -1 and t1.row_nm + 1