I need to store data with varchar name and Integer intValue. All integer values are unique and I need to keep up that contract
I need to write query to add the element using the following rule: if after insertion there is an intValue duplication - we need to increase intValue of existed element to resolve conflict. Repeat that operation until no conflict left.
Example:
B | 2 | | B | 2 |
C | 3 | | E | 3 |
D | 4 | => insert (E 3) => | C | 4 |
A | 1 | | D | 5 |
Z | 7 | | A | 1 |
| Z | 7 |
The only idea is to run update query in a loop but that looks too unefficient.
I need to write this query in Spring JPA, so the only requirement that the query should not be database specific
Business case:
Let's say there is a people in the queue. And intValue is position in the queue. So, "Add" means that some person come, pay money and say: I dont wanna be the last in the queue. I want to be, for example, the 3rd. So you take the money and put that person in a queue so other people after him - increments their position.
The only difference from the queue - that in my case there are gaps allowed
Aha, we might say that the gaps are occasioned by people leaving the queue.
Lets try this. Loops are inevitable--either server does them, or we can do as SQL.
-- prepare test data
declare #PeopleQueue table (pqname varchar(100), intValue int);
insert into #PeopleQueue
SELECT 'B' AS pqname, 2 as intValue UNION ALL
SELECT 'C' AS pqname, 3 as intValue UNION ALL
SELECT 'D' AS pqname, 4 as intValue UNION ALL
SELECT 'A' AS pqname, 1 as intValue UNION ALL
SELECT 'Z' AS pqname, 7 as intValue
;
--SELECT '' AS pqname, 0 as intValue UNION ALL
Select * from #PeopleQueue; - verify good test data
-- Solve the problem
Declare #pqnameNEW varchar(100) = 'E';
Declare #intNEW int = 3; -- 3 for conflict, or for no conflict, use 13
Declare #intHIGH int;
IF EXISTS ( SELECT 1 FROM #PeopleQueue WHERE intValue = #intNEW )
BEGIN
-- find the end of the sequence, before the gap
SET #intHIGH = (
SELECT TOP 1
intValue
FROM #PeopleQueue pq
WHERE NOT EXISTS
(
SELECT NULL
FROM #PeopleQueue pn
WHERE pn.intValue = pq.intValue + 1
)
AND pq.intValue >= #intNEW
)
;
-- now Update all from intNEW thru intHIGH
UPDATE #PeopleQueue
SET intValue = intValue + 1
WHERE intValue >= #intNEW
AND intValue <= #intHIGH
End;
-- finally insert the new item
INSERT into #PeopleQueue Values (#pqnameNEW, #intNEW);
Select * from #PeopleQueue; -- verify correct solution
Edited--11/28 17:00
Or, estimate the number of Bump-the-Line-Inserts (vs append to the end inserts), and design the intValues to be originally in multiples of ten (10) so that long sequences of updates are minimized.
update queue
SET intValue = intValue + 1
WHERE intValue >= 3
AND intValue <= (
SELECT q1.intValue
FROM queue as q1 LEFT JOIN queue AS q2 ON q1.intValue + 1 = q2.intValue
WHERE q2.name is NULL AND q1.intValue > 3
ORDER BY q1.intValue
LIMIT 1
)
My table called element looks like this:
id | successor | important
----------------------------
1 | NULL | 0
2 | 4 | 1
3 | 5 | 0
4 | 8 | 0
5 | 6 | 1
6 | 7 | 0
7 | NULL | 0
8 | 10 | 1
9 | 10 | 0
10 | NULL | 0
I start with an element’s ID. Each element may or may not have a succeeding element. So given any element ID I may build a chain of elements from 0..n elements depending on its successors and successor-successors, and so on.
Let’s say my starting ID is 2. This results in the following chain:
2 -> 4 -> 8 -> 10
Now I want to ask this question: Does a specific element chain contain at least one element where important == 1?
In pseudo-code a function realizing this without unneccessary checks may look like this:
boolean chainIsImportant(element)
{
if (element.important == 1) {
return true;
}
if (element.successor != NULL) {
return chainIsImportant(element.successor);
}
return false;
}
I guess this can be realized with WITH RECURSIVE, right? How can I stop recursion, once an element with important == 1 was found?
This is typically done by aggregating the columns in question and adding a condition on the join in the recursive part of the CTE:
with recursive all_elements as (
select id, successor, important, array[important] as important_elements
from elements
where successor is null
union all
select c.id, c.successor, c.important, p.important_elements||c.important
from elements c
join all_elements p on c.successor = p.id
where 1 <> all(p.important_elements)
)
select *
from all_elements;
Note that the condition is "flipped" because the where clause defines those rows that should be included.
CREATE TABLE booltree
( id INTEGER NOT NULL PRIMARY KEY
, successor INTEGER REFERENCES booltree(id)
, important Boolean NOT NULL
);
INSERT INTO booltree(id , successor , important) VALUES
( 1, NULL , False)
,(2, 4 , True)
,(3, 5 , False)
,(4, 8 , False)
,(5, 6 , True)
,(6, 7 , False)
,(7, NULL , False)
,(8, 10 , True)
,(9, 10 , False)
,(10, NULL , False)
;
-- SELECT * FROM booltree;
WITH RECURSIVE rec AS (
SELECT id, important
FROM booltree
WHERE successor IS NULL
UNION ALL
SELECT bt.id, GREATEST(rec.important, bt.important) AS important
FROM booltree bt
JOIN rec ON bt.successor = rec.id
)
SELECT id, important
FROM rec
ORDER BY important, id;
Result:
CREATE TABLE
INSERT 0 10
id | important
----+-----------
1 | f
6 | f
7 | f
9 | f
10 | f
2 | t
3 | t
4 | t
5 | t
8 | t
(10 rows)
Note: IMHO the recursion cannot be stopped once a True importance is found (basically, because LEFT JOINS are not allowed in RECURSIVE UNIONS)
But if you are looking for exactly one given id (or a set of them) then maybe you could use that as the start condition, and search the tree upwards.
I'm trying to merge tables where rows correspond to a many:1 relationship with "real" things.
I'm writing a blackjack simulator that stores game history in a database with a new set of tables generated each run. The tables are really more like templates, since each game gets its own set of the 3 mutable tables (players, hands, and matches). Here's the layout, where suff is a user-specified suffix to use for the current run:
- cards
- id INTEGER PRIMARY KEY
- cardValue INTEGER NOT NULL
- suit INTEGER NOT NULL
- players_suff
- whichPlayer INTEGER PRIMARY KEY
- aiType TEXT NOT NULL
- hands_suff
- id BIGSERIAL PRIMARY KEY
- whichPlayer INTEGER REFERENCES players_suff(whichPlayer) *
- whichHand BIGINT NOT NULL
- thisCard INTEGER REFERENCES cards(id)
- matches_suff
- id BIGSERIAL PRIMARY KEY
- whichGame INTEGER NOT NULL
- dealersHand BIGINT NOT NULL
- whichPlayer INTEGER REFERENCES players_suff(whichPlayer)
- thisPlayersHand BIGINT NOT NULL **
- playerResult INTEGER NOT NULL --AKA who won
Only one cards table is created because its values are constant.
So after running the simulator twice you might have:
hands_firstrun
players_firstrun
matches_firstrun
hands_secondrun
players_secondrun
matches_secondrun
I want to be able to combine these tables if you used the same AI parameters for both of those runs (i.e. players_firstrun and players_secondrun are exactly the same). The problem is that the way I'm inserting hands makes this really messy: whichHand can't be a BIGSERIAL because the relationship of hands_suff rows to "actual hands" is many:1. matches_suff is handled the same way because a blackjack "game" actually consists of a set of games: the set of pairs of each player vs. the dealer. So for 3 players, you actually have 3 rows for each round.
Currently I select the largest whichHand in the table, add 1 to it, then insert all of the rows for one hand. I'm worried this "query-and-insert" will be really slow if I'm merging 2 tables that might both be arbitrarily huge.
When I'm merging tables, I feel like I should be able to (entirely in SQL) query the largest values in whichHand and whichGame once then use them combine the tables, incrementing them for each unique whichHand and whichGame in the table being merged.
(I saw this question, but it doesn't handle using a generated ID in 2 different places). I'm using Postgres and it's OK if the answer is specific to it.
* sadly postgres doesn't allow parameterized table names so this had to be done by manual string substitution. Not the end of the world since the program isn't web-facing and no one except me is likely to ever bother with it, but the SQL injection vulnerability does not make me happy.
** matches_suff(whichPlayersHand) was originally going to reference hands_suff(whichHand) but foreign keys must reference unique values. whichHand isn't unique because a hand is made up of multiple rows, with each row "holding" one card. To query for a hand you select all of those rows with the same value in whichHand. I couldn't think of a more elegant way to do this without resorting to arrays.
EDIT:
This is what I have now:
thomas=# \dt
List of relations
Schema | Name | Type | Owner
--------+----------------+-------+--------
public | cards | table | thomas
public | hands_first | table | thomas
public | hands_second | table | thomas
public | matches_first | table | thomas
public | matches_second | table | thomas
public | players_first | table | thomas
public | players_second | table | thomas
(7 rows)
thomas=# SELECT * FROM hands_first
thomas-# \g
id | whichplayer | whichhand | thiscard
----+-------------+-----------+----------
1 | 0 | 0 | 6
2 | 0 | 0 | 63
3 | 0 | 0 | 41
4 | 1 | 1 | 76
5 | 1 | 1 | 23
6 | 0 | 2 | 51
7 | 0 | 2 | 29
8 | 0 | 2 | 2
9 | 0 | 2 | 92
10 | 0 | 2 | 6
11 | 1 | 3 | 101
12 | 1 | 3 | 8
(12 rows)
thomas=# SELECT * FROM hands_second
thomas-# \g
id | whichplayer | whichhand | thiscard
----+-------------+-----------+----------
1 | 0 | 0 | 78
2 | 0 | 0 | 38
3 | 1 | 1 | 24
4 | 1 | 1 | 18
5 | 1 | 1 | 95
6 | 1 | 1 | 40
7 | 0 | 2 | 13
8 | 0 | 2 | 84
9 | 0 | 2 | 41
10 | 1 | 3 | 29
11 | 1 | 3 | 34
12 | 1 | 3 | 56
13 | 1 | 3 | 52
thomas=# SELECT * FROM matches_first
thomas-# \g
id | whichgame | dealershand | whichplayer | thisplayershand | playerresult
----+-----------+-------------+-------------+-----------------+--------------
1 | 0 | 0 | 1 | 1 | 1
2 | 1 | 2 | 1 | 3 | 2
(2 rows)
thomas=# SELECT * FROM matches_second
thomas-# \g
id | whichgame | dealershand | whichplayer | thisplayershand | playerresult
----+-----------+-------------+-------------+-----------------+--------------
1 | 0 | 0 | 1 | 1 | 0
2 | 1 | 2 | 1 | 3 | 2
(2 rows)
I'd like to combine them to have:
hands_combined table:
id | whichplayer | whichhand | thiscard
----+-------------+-----------+----------
1 | 0 | 0 | 6 --Seven of Spades
2 | 0 | 0 | 63 --Queen of Spades
3 | 0 | 0 | 41 --Three of Clubs
4 | 1 | 1 | 76
5 | 1 | 1 | 23
6 | 0 | 2 | 51
7 | 0 | 2 | 29
8 | 0 | 2 | 2
9 | 0 | 2 | 92
10 | 0 | 2 | 6
11 | 1 | 3 | 101
12 | 1 | 3 | 8
13 | 0 | 4 | 78
14 | 0 | 4 | 38
15 | 1 | 5 | 24
16 | 1 | 5 | 18
17 | 1 | 5 | 95
18 | 1 | 5 | 40
19 | 0 | 6 | 13
20 | 0 | 6 | 84
21 | 0 | 6 | 41
22 | 1 | 7 | 29
23 | 1 | 7 | 34
24 | 1 | 7 | 56
25 | 1 | 7 | 52
matches_combined table:
id | whichgame | dealershand | whichplayer | thisplayershand | playerresult
----+-----------+-------------+-------------+-----------------+--------------
1 | 0 | 0 | 1 | 1 | 1
2 | 1 | 2 | 1 | 3 | 2
3 | 2 | 4 | 1 | 5 | 0
4 | 3 | 6 | 1 | 7 | 2
Each value of "thiscard" represents a playing card in the range [1..104]--52 playing cards with an extra bit representing if it's face up or face down. I didn't post the actual table for space reasons.
So player 0 (aka the dealer) had a hand of (Seven of Spades, Queen of Spaces, 3 of Clubs) in the first game.
I think you're not using PostgreSQL the way it's intended to be used, plus your table design may not be suitable for what you want to achieve. Whilst it was difficult to understand what you want your solution to achieve, I wrote this, which seems to solve everything you want using a handful of tables only, and functions that return recordsets for simulating your requirement for individual runs. I used Enums and complex types to illustrate some of the features that you may wish to harness from the power of PostgreSQL.
Also, I'm not sure what parameterized table names are (I have never seen anything like it in any RDBMS), but PostgreSQL does allow something perfectly suitable: recordset returning functions.
CREATE TYPE card_value AS ENUM ('1', '2', '3', '4', '5', '6', '7', '8', '9', '10', 'J', 'Q', 'K');
CREATE TYPE card_suit AS ENUM ('Clubs', 'Diamonds', 'Hearts', 'Spades');
CREATE TYPE card AS (value card_value, suit card_suit, face_up bool);
CREATE TABLE runs (
run_id bigserial NOT NULL PRIMARY KEY,
run_date timestamptz NOT NULL DEFAULT CURRENT_TIMESTAMP
);
CREATE TABLE players (
run_id bigint NOT NULL REFERENCES runs,
player_no int NOT NULL, -- 0 can be assumed as always the dealer
ai_type text NOT NULL,
PRIMARY KEY (run_id, player_no)
);
CREATE TABLE matches (
run_id bigint NOT NULL REFERENCES runs,
match_no int NOT NULL,
PRIMARY KEY (run_id, match_no)
);
CREATE TABLE hands (
hand_id bigserial NOT NULL PRIMARY KEY,
run_id bigint NOT NULL REFERENCES runs,
match_no int NOT NULL,
hand_no int NOT NULL,
player_no int NOT NULL,
UNIQUE (run_id, match_no, hand_no),
FOREIGN KEY (run_id, match_no) REFERENCES matches,
FOREIGN KEY (run_id, player_no) REFERENCES players
);
CREATE TABLE deals (
deal_id bigserial NOT NULL PRIMARY KEY,
hand_id bigint NOT NULL REFERENCES hands,
card card NOT NULL
);
CREATE OR REPLACE FUNCTION players(int) RETURNS SETOF players AS $$
SELECT * FROM players WHERE run_id = $1 ORDER BY player_no;
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION matches(int) RETURNS SETOF matches AS $$
SELECT * FROM matches WHERE run_id = $1 ORDER BY match_no;
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION hands(int) RETURNS SETOF hands AS $$
SELECT * FROM hands WHERE run_id = $1 ORDER BY match_no, hand_no;
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION hands(int, int) RETURNS SETOF hands AS $$
SELECT * FROM hands WHERE run_id = $1 AND match_no = $2 ORDER BY hand_no;
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION winner_player (int, int) RETURNS int AS $$
SELECT player_no
FROM hands
WHERE run_id = $1 AND match_no = $2
ORDER BY hand_no DESC
LIMIT 1
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION next_player_no (int) RETURNS int AS $$
SELECT CASE WHEN EXISTS (SELECT 1 FROM runs WHERE run_id = $1) THEN
COALESCE((SELECT MAX(player_no) FROM players WHERE run_id = $1), 0) + 1 END
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION next_match_no (int) RETURNS int AS $$
SELECT CASE WHEN EXISTS (SELECT 1 FROM runs WHERE run_id = $1) THEN
COALESCE((SELECT MAX(match_no) FROM matches WHERE run_id = $1), 0) + 1 END
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION next_hand_no (int) RETURNS int AS $$
SELECT CASE WHEN EXISTS (SELECT 1 FROM runs WHERE run_id = $1) THEN
COALESCE((SELECT MAX(hand_no) + 1 FROM hands WHERE run_id = $1), 0) END
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION card_to_int (card) RETURNS int AS $$
SELECT ((SELECT enumsortorder::int-1 FROM pg_enum WHERE enumtypid = 'card_suit'::regtype AND enumlabel = ($1).suit::name) * 13 +
(SELECT enumsortorder::int-1 FROM pg_enum WHERE enumtypid = 'card_value'::regtype AND enumlabel = ($1).value::name) + 1) *
CASE WHEN ($1).face_up THEN 2 ELSE 1 END
$$ LANGUAGE SQL; -- SELECT card_to_int(('3', 'Spades', false))
CREATE OR REPLACE FUNCTION int_to_card (int) RETURNS card AS $$
SELECT ((SELECT enumlabel::card_value FROM pg_enum WHERE enumtypid = 'card_value'::regtype AND enumsortorder = ((($1-1)%13)+1)::real),
(SELECT enumlabel::card_suit FROM pg_enum WHERE enumtypid = 'card_suit'::regtype AND enumsortorder = (((($1-1)/13)::int%4)+1)::real),
$1 > (13*4))::card
$$ LANGUAGE SQL; -- SELECT i, int_to_card(i) FROM generate_series(1, 13*4*2) i
CREATE OR REPLACE FUNCTION deal_cards(int, int, int, int[]) RETURNS TABLE (player_no int, hand_no int, card card) AS $$
WITH
hand AS (
INSERT INTO hands (run_id, match_no, player_no, hand_no)
VALUES ($1, $2, $3, next_hand_no($1))
RETURNING hand_id, player_no, hand_no),
mydeals AS (
INSERT INTO deals (hand_id, card)
SELECT hand_id, int_to_card(card_id)::card AS card
FROM hand, UNNEST($4) card_id
RETURNING hand_id, deal_id, card
)
SELECT h.player_no, h.hand_no, d.card
FROM hand h, mydeals d
$$ LANGUAGE SQL;
CREATE OR REPLACE FUNCTION deals(int) RETURNS TABLE (deal_id bigint, hand_no int, player_no int, card int) AS $$
SELECT d.deal_id, h.hand_no, h.player_no, card_to_int(d.card)
FROM hands h
JOIN deals d ON (d.hand_id = h.hand_id)
WHERE h.run_id = $1
ORDER BY d.deal_id;
$$ LANGUAGE SQL;
INSERT INTO runs DEFAULT VALUES; -- Add first run
INSERT INTO players VALUES (1, 0, 'Dealer'); -- dealer always zero
INSERT INTO players VALUES (1, next_player_no(1), 'Player 1');
INSERT INTO matches VALUES (1, next_match_no(1)); -- First match
SELECT * FROM deal_cards(1, 1, 0, ARRAY[6, 63, 41]);
SELECT * FROM deal_cards(1, 1, 1, ARRAY[76, 23]);
SELECT * FROM deal_cards(1, 1, 0, ARRAY[51, 29, 2, 92, 6]);
SELECT * FROM deal_cards(1, 1, 1, ARRAY[101, 8]);
INSERT INTO matches VALUES (1, next_match_no(1)); -- Second match
SELECT * FROM deal_cards(1, 2, 0, ARRAY[78, 38]);
SELECT * FROM deal_cards(1, 2, 1, ARRAY[24, 18, 95, 40]);
SELECT * FROM deal_cards(1, 2, 0, ARRAY[13, 84, 41]);
SELECT * FROM deal_cards(1, 2, 1, ARRAY[29, 34, 56, 52]);
SELECT * FROM deals(1); -- This is the output you need (hands_combined table)
-- This view can be used to retrieve the list of all winning hands
CREATE OR REPLACE VIEW winning_hands AS
SELECT DISTINCT ON (run_id, match_no) *
FROM hands
ORDER BY run_id, match_no, hand_no DESC;
SELECT * FROM winning_hands;
Wouldn't using the UNION operator work?
For the hands relation:
SELECT * FROM hands_first
UNION ALL
SELECT * FROM hands_second
For the matches relation:
SELECT * FROM matches_first
UNION ALL
SELECT * FROM matches_second
As a more long term solution I'd consider restructuring the DB because it will quickly become unmanageable with this schema. Why not improve normalization by introducing a games table?
In other words Games have many Matches, matches have many players for each game and players have many hands for each match.
I'd recommend drawing the UML for the entity relationships on paper (http://dawgsquad.googlecode.com/hg/docs/database_images/Database_Model_Diagram(Title).png), then improving the schema so it can be queried using normal SQL operators.
Hope this helps.
EDIT:
In that case you can use a subquery on the union of both tables with the rownumber() PG function to represent the row number:
SELECT
row_number() AS id,
whichplayer,
whichhand,
thiscard
FROM
(
SELECT * FROM hands_first
UNION ALL
SELECT * FROM hands_second
);
The same principle would apply to the matches table. Obviously this doesn't scale well to even a small number of tables, so would prioritize normalizing your schema.
Docs on some PG functions: http://www.postgresql.org/docs/current/interactive/functions-window.html
to build new table with all rows of two tables, do:
CREATE TABLE hands AS
select 1 as hand, id, whichplayer, whichhand, thiscard
from hands_first
union all
select 2 as hand, id, whichplayer, whichhand, thiscard
from hands_second
after that, to insert data of new matche, create sequence with start on current last + 1
CREATE SEQUENCE matche START 3;
before insert read sequence value, and use it in inserts:
SELECT nextval('matche');
Your database structure is not great, and I know for sure it is not scalable approach creating tables on fly. There are performance drawbacks creating physical tables instead of using an existing structure. I suggest you refactor your db structure if can.
You can however use the UNION operator to merge your data.
I have to handle a table PRODUCTS which is created to accommodate tree structure of products. It is done to handle situations when one product can contain several others (e.g. one package product holds several other positions). So, I'm making a function that takes OrderDetails, and it must iterate through all PRODUCTS and list out the child products for each product listed. I am facing an issue that I have to iterate through tree of unknown depth. Please, give me an idea how to do it.
I've implemented it in the table below with the function listed along with it. But in that solution the depth of listing is limited to 1, and what i want to do is to fetch all depth of the tree.
Here is the code:
CREATE OR REPLACE FUNCTION foo()RETURNS text AS
$body$
DECLARE _row RECORD;
_result text := '';
_child_row RECORD;
_count integer := -1;
_marker integer := 1;
BEGIN
FOR _row IN SELECT * FROM tree_products
LOOP
_result := _result || _marker || ' ' || _row.name;
_count := (SELECT count(product_id) FROM tree_products WHERE parent_id = _row.product_id);
IF _count > 0 THEN
FOR _child_row IN SELECT * FROM tree_products WHERE parent_id = _row.product_id
LOOP
_result := _result || ' ' || _child_row.name;
END LOOP;
END IF;
_marker := _marker =1;
END LOOP;
END;
$body$
LANGUAGE plpgsql
UPD Done this usign WITH CTE, but the groupiing problem occured:
CREATE OR REPLACE FUNCTION public.__foo (
)
RETURNS SETOF refcursor AS
$body$
DECLARE _returnvalue refcursor;
_q text;
BEGIN
_q :='
WITH RECURSIVE r_p (product_id, name, parent_id) AS -- 1
(SELECT t_p.product_id, t_p.name , t_p.parent_id -- 2
FROM tree_products t_p
WHERE t_p.product_id = 1
UNION ALL
SELECT t_c.product_id, t_c.name, t_c.parent_id -- 3
FROM r_p t_p, tree_products t_c
WHERE t_c.parent_id = t_p.product_id)
SELECT product_id, name, parent_id -- 4
FROM r_p;';
OPEN _returnvalue FOR EXECUTE (_q);
RETURN NEXT _returnvalue;
END
$body$
LANGUAGE 'plpgsql'
VOLATILE
CALLED ON NULL INPUT
SECURITY INVOKER
COST 100 ROWS 1000;
I want to sibling products be under their respectiveparents, I wonder how to write Grouping statement...
UPD Sorry, the definition of the tree_products is the following:
CREATE TABLE public.tree_products (
product_id INTEGER DEFAULT nextval('ree_products_product_id_seq'::regclass) NOT NULL,
name VARCHAR,
parent_id INTEGER,
CONSTRAINT ree_products_pkey PRIMARY KEY(product_id)
)
WITH (oids = false);
UPD: SAMPLE OUTPUT:
product_id | name | parent_id
---------------------------------------
1 | promo | NULL
3 | fork | 1
4 | spoon | 1
6 | can | 1
10 | big can | 3
11 | small can | 4
12 | large spoon | 6
13 | mega fork | 3
14 | super duper | 6
DESIRED OUTPUT:
product_id | name | parent_id
---------------------------------------
1 | promo | NULL
3 | fork | 1
10 | big can | 3
13 | mega fork | 3
4 | spoon | 1
11 | small can | 4
6 | can | 1
12 | large spoon | 6
14 | super duper | 6
So, the fetched table has structure of the real tree, like the follwing:
- promo
- fork
- big can
- mega fork
- spoon
- small can
- can
- large can
- super duper
This SQLFiddle traverses the tree top-down, keeping an list of parent row numbers in an array, essentially a "parent row position list".
It then sorts the results by the parent-list.
WITH RECURSIVE tree(product_id, name, parentlist) AS (
SELECT product_id, name, ARRAY[ row_number() OVER (ORDER BY product_id) ]
FROM tree_products
WHERE parent_id IS NULL
UNION
SELECT tp.product_id, tp.name, array_append(parentlist, row_number() OVER (ORDER BY tp.product_id))
FROM tree_products tp
INNER JOIN tree t
ON (tp.parent_id = t.product_id)
)
SELECT *
FROM tree
ORDER BY parentlist;
I have 2 tables: errorlookup and errors.
errorlookup has 2 columns: codes and description.
The codes are of length 2.
errors has 2 columns id and errorcodes.
The errorcodes are of length 40 meaning they code store 20 error codes for each id.
I need to display all the description associated with the id by substring the errorcodes and matching with code in errorlookup table.
Sample data for errorlookup:
codes:description
12:Invalid
22:Inactive
21:active
Sample data for errors:
id:errorcodes
1:1221
2:2112
3:1222
I cant use LIKE as it would result in too many errors. I want the errorcodes column to be broken down into strings of length 2 and then joined with the errorlookup.
How can it be done?
If you really cannot alter the tables structure, here's another approach:
Create an auxilary numbers table:
CREATE TABLE numbers
( i INT NOT NULL
, PRIMARY KEY (i)
) ;
INSERT INTO numbers VALUES
( 1 ) ;
INSERT INTO numbers VALUES
( 2 ) ;
--- ...
INSERT INTO numbers VALUES
( 100 ) ;
Then you could use this:
SELECT err.id
, err.errorcodes
, num.i
, look.codes
, look.descriptionid
FROM
( SELECT i, 2*i-1 AS pos --- odd numbers
FROM numbers
WHERE i <= 20 --- 20 pairs
) num
CROSS JOIN
errors err
JOIN
errorlookup look
ON look.codes = SUBSTR(err.errorcodes, pos, 2)
ORDER BY
err.errorcodes
, num.i ;
Test at: SQL-Fiddle
ID ERRORCODES I CODES DESCRIPTIONID
1 1221 1 12 Invalid
1 1221 2 21 Active
3 1222 1 12 Invalid
3 1222 2 22 Inactive
2 2112 1 21 Active
2 2112 2 12 Invalid
I think the cleanest solution is to "normalize" your errocodes table using a PL/SQL function. That way you can keep the current (broken) table design, but still access its content as if it was properly normlized.
create type error_code_type as object (id integer, code varchar(2))
/
create or replace type error_table as table of error_code_type
/
create or replace function unnest_errors
return error_table pipelined
is
codes_l integer;
i integer;
one_row error_code_type := error_code_type(null, null);
begin
for err_rec in (select id, errorcodes from errors) loop
codes_l := length(err_rec.errorcodes);
i := 1;
while i < codes_l loop
one_row.id := err_rec.id;
one_row.code := substr(err_rec.errorcodes, i, 2);
pipe row (one_row);
i := i + 2;
end loop;
end loop;
end;
/
Now with this function you can do something like this:
select er.id, er.code, el.description
from table(unnest_errors) er
join errorlookup el on el.codes = er.code;
You can also create a view based on the function to make the statements a bit easier to read:
create or replace view normalized_errorcodes
as
select *
from table(unnest_errors);
Then you can simply reference the view in the real statement.
(I tested this on 11.2 but I believe it should work on 10.x as well)
I think you're on the right track with LIKE. MySQL has an RLIKE function that allows matching by regular expression (I don't know if it's present in Oracle.) You could use errorlookup.code as a pattern to match against errors.errorcodes. The (..)* pattern is used to prevent things like "1213" from matching, for example, "21".
SELECT *
FROM error
JOIN errorlookup
WHERE errorcodes RLIKE CONCAT('^(..)*',code)
ORDER BY id;
+------+----------+------+
| id | errorcode| code |
+------+----------+------+
| 1 | 11 | 11 |
| 2 | 1121 | 11 |
| 2 | 1121 | 21 |
| 3 | 21313245 | 21 |
| 3 | 21313245 | 31 |
| 3 | 21313245 | 32 |
| 4 | 21 | 21 |
+------+----------+------+