When I run this query in my database I get the following:
WITH RECURSIVE series AS (
SELECT CONCAT( a.title) as str, a.prequelID
FROM ( Prequels NATURAL JOIN Books) AS a
UNION
SELECT CONCAT(t.title, ' -> ', str) as str, t.prequelID
FROM (Books NATURAL JOIN Prequels) as t
INNER JOIN series AS s ON s.prequelID = t.bookID
)
SELECT str as series FROM series
ORDER BY series;
This is the result :
I don't want any duplicates, only the full string that shows the whole series. How do I do this?
UPDATE:
I updated the query because I realised I was missing the first book in the series.
WITH RECURSIVE series AS (
SELECT CONCAT( a.title) as str, a.prequelID
FROM ( Prequels NATURAL JOIN Books) AS a
UNION
SELECT CONCAT(t.title, ' -> ', str) as str, t.prequelID
FROM (Books NATURAL JOIN Prequels) as t
INNER JOIN series AS s ON s.prequelID = t.bookID
)
SELECT CONCAT(a.title, ' -> ',a.str) as series
FROM (
SELECT Books.title, series.str
FROM series JOIN Books ON series.prequelID = Books.bookID
WHERE NOT EXISTS(
SELECT prequelID
FROM Prequels
WHERE series.prequelID = Prequels.bookID
)
) a
ORDER BY series;
The result is still a bit off, since I only want the complete string of the series:
How do I fix this?
Tables:
CREATE TABLE Books
(bookID integer PRIMARY KEY,
title varchar(100),
pages integer);
CREATE TABLE Prequels
(bookID INTEGER REFERENCES Books(bookID),
prequelID INTEGER REFERENCES Books(bookID),
PRIMARY KEY (bookID,prequelID));
Sample data of the Game of Thornes series:
INSERT INTO BOOKS (bookID,title,pages) VALUES (80429,'A Game of Thrones',292);
INSERT INTO BOOKS (bookID,title,pages) VALUES (41121,'A Clash of Kings',160);
INSERT INTO BOOKS (bookID,title,pages) VALUES (29287,'A Storm of Swords',160);
INSERT INTO BOOKS (bookID,title,pages) VALUES (17696,'A Feast for Crows',292);
INSERT INTO BOOKS (bookID,title,pages) VALUES (3947,'A Dance with Dragons',101);
INSERT INTO Prequels (bookID,prequelID) VALUES (41121,80429);
INSERT INTO Prequels (bookID,prequelID) VALUES (29287,41121);
INSERT INTO Prequels (bookID,prequelID) VALUES (17696,29287);
INSERT INTO Prequels (bookID,prequelID) VALUES (3947,17696);
CodePudding user response:
You'd want to start from the roots of the trees.
There are no roots in the prequelId's.
WITH RECURSIVE series AS (
SELECT
b.bookId
, 1 as lvl
, p.prequelID
, CONCAT(b.title) as series
FROM Books AS b
LEFT JOIN Prequels AS p
ON p.bookId = b.bookId
WHERE NOT EXISTS (
SELECT 1
FROM Prequels p2
WHERE p2.prequelID = b.bookId
)
UNION ALL
SELECT
s.bookId
, s.lvl 1
, p.prequelID
, CONCAT(b.title, ' -> ', s.series)
FROM series AS s
JOIN Books AS b
ON b.bookId = s.prequelID
LEFT JOIN Prequels AS p
ON p.bookId = s.prequelID
)
SELECT series
FROM series
WHERE prequelId is null
AND lvl > 1
ORDER BY series;
series |
---|
A Game of Thrones -> A Clash of Kings -> A Storm of Swords -> A Feast for Crows -> A Dance with Dragons |
Demo on db<>fiddle here
CodePudding user response:
Modified version of LukStorms now deleted answer...
https://dbfiddle.uk/?rdbms=postgres_14&fiddle=4e26beed430aefb60f1dc91766db13d3
WITH RECURSIVE series AS (
SELECT
b.bookId,
1 as lvl,
p.prequelID,
CONCAT(b.title) AS series
FROM
books b
LEFT JOIN
prequels p
ON p.bookid = b.bookid
WHERE NOT EXISTS (
SELECT 1
FROM Prequels p
WHERE p.prequelID = b.bookId
)
UNION ALL
SELECT
b.bookid,
lvl 1,
p.prequelID,
CONCAT(b.title, ' -> ', s.series)
FROM
series s
INNER JOIN
books b
ON b.bookid = s.prequelid
LEFT JOIN
prequels p
ON p.bookid = b.bookid
)
SELECT
s.series
FROM
series s
WHERE
s.prequelid IS NULL
ORDER BY
s.series;
The non recuraive part of the CTE has a LEFT JOIN so to include books that are not part of a series, just a standalone book. Revert that to NATURAL or INNER JOIN if you don't want that.
The non recursive part of the CTE has a WHERE clause to ensure it only ever starts from the last book in a series.
Thr recursive part of the CTE has a LEFT JOIN on the prequels table, so that the first book in a series isn't lost (due to not having a prequel row), and so records that book's prequel as NULL.
The outer query then looks for that NULL, so only outputting rows for completed series.
CodePudding user response:
After the recursive query, you can compare all the str together with a self join, and keep the s1.str
which are not a substring of any s2.str
thanks to the WHERE s2.str IS NULL
clause.
WITH RECURSIVE series AS (
SELECT CONCAT( a.title) as str, a.prequelID
FROM ( Prequels NATURAL JOIN Books) AS a
UNION
SELECT CONCAT(t.title, ' -> ', str) as str, t.prequelID
FROM (Books NATURAL JOIN Prequels) as t
INNER JOIN series AS s ON s.prequelID = t.bookID
)
SELECT s1.str as series
FROM series AS s1
LEFT JOIN series AS s2
ON s2.str ~ s1.str
AND s2.str <> s1.str
WHERE s2.str IS NULL
ORDER BY series;
test result in dbfiddle