I have a test table with columns(id, title, language)
id: primary_key (auto-increment)
unique_key(composite) on language and title
id | title | language |
---|---|---|
1 | Japanese | JP |
2 | Australian | AU |
3 | English | EN |
4 | Hindi | HI |
I would like to have a query which either return 1 or 0 rows based on the multiple language criteria.
The query should return the result with language priority from left to right, if no row found for first language from left then look for the second language as so on.
Use-case:
languages | result row_id | remarks |
---|---|---|
JP,HI | 1 | As JP found at id=1 |
HI,JP | 4 | As HI found at id=4 |
RU,AU | 2 | As AU found at id=2, no row with language=RU |
PK,ST | no row | As no language has value PK or ST |
Here's I've tried custom order using FIELD
clause on language column:
SELECT id, title, language
FROM test WHERE language IN ('TH', 'AU','EN') ORDER BY
FIELD(LANGUAGE,'TH','AU','EN') LIMIT 1;
Expected/Actual output:
id | title | language |
---|---|---|
2 | Australian | AU |
I would like to know is there any better way(in terms of performance and readability) to achieve this use-case?
CodePudding user response:
Convert languages list to rowset:
SELECT test.id, test.title, language
FROM test
JOIN (SELECT 'TH' language, 1 priority UNION ALL
SELECT 'AU' , 2 UNION ALL
SELECT 'EN' , 3 ) languages USING (language)
ORDER BY languages.priority LIMIT 1;
CodePudding user response:
See if this will help:
SELECT ...
WHERE FIND_IN_SET(language, languages)
ORDER BY FIND_IN_SET(language, languages)
FIND_IN_SET will return the empty set if there is no match, so something extra is needed to handle the "0" case.