Question

Let’s say I have table with column ‘URL’ whrere I store urls like this

• one/two

• one/two/three

• alpha/omega

And I want to get data from database for specific url and if it is not found I remove the last part of url and search again:

Example:

I have url like one/two/three/four/five.

I do search for “one/two/three/four/five”

if not found search again for “one/two/three/four”

if not found search again for “one/two/three”

if not found search again for “one/two”

I would like to have something like:

SELECT * FROM db WHERE url=one/two/three/four/five

UNION

SELECT * FROM db WHERE url=one/two/three/four/five

UNION

SELECT * FROM db WHERE url=one/two/three/four

UNION

SELECT * FROM db WHERE url=one/two/three

UNION

SELECT * FROM db WHERE url=one/two

UNION

SELECT * FROM db WHERE url=one

but I want to stop searching if the row is found.

Is this possible or do I have to do it with separated queries.

Thanks for help.

Answer 1

I thing that this is the most elegant approach to your question. This statement is independent depth path and you don’t need to split constant url in subsequent selects:

SELECT 
   * 
FROM 
   db 
WHERE 
   concat( 'one/two/three/four/five' , '/') like concat( url , '/%')
ORDER BY
   LENGTH (url) desc
LIMIT 1

I have tested this query in MySQL, also you can check it! (in MSSQL syntax)