Question

I have a request of translation:
(table structure:

1. line_id
2. lang (“en” or “ru”)
3. text

so, in the table it looks like:

5 | en | Test
6 | en | Hello!
7 | en | Another words
6 | ru | Привет!

and all translations are in the same table with key line_id, that is independent from the language)

SELECT txt.line_id, txt.text AS o, ru.text AS t 
  FROM text AS txt, text AS ru 
 WHERE txt.line_id = ru.line_id 
   AND txt.lang = 'en';

it will return such array

> [5] => Array ( [line_id] => 5 [o] => Test [t] => Test 
> [6] => Array ( [line_id] => 6 [o] => Hello! [t] => Hello! ) 
> [7] => Array ( [line_id] => 6 [o] => Hello! [t] => Привет! )

o – is original text, t – translation.

How to delete from set #6 row, because we have translation in the next row. GROUP BY will kill #7 and save #6 row.

the best result would be:

> [5] => Array ( [line_id] => 5 [o] => Test [t] => ) 
> [6] => Array ( [line_id] => 6 [o] => Hello! [t] => Привет! )

without [o] => Hello! [t] => Hello!

Answer 1

I’m just guessing you probably need:

SELECT txt.line_id, 
         txt.text AS o, 
         COALESCE (ru.text,'UNTRANSLATED') AS t 
FROM 
  text AS txt
LEFT JOIN 
  text AS ru 
ON txt.line_id=ru.line_id AND ru.lang='ru'
WHERE txt.lang='en'