Select variable amount of columns and htp.print them

Question

Hi,
i have a rather difficult problem i guess. I have a table "Languages" in which i have a couple of languages for example 'nl', 'f', 'eg', ...
i also have another table "Translations", which consists of an id and variable columns. For example, if we have the languages 'nl','f' and 'eg' in our Languages table, we also have a corresponding column in the Translations column. The column names are then 'nl', 'f' and 'eg'. (i'm using nls).
So if i add a language, two things happen:
-a row is added to the languages table
-the translations table is altered; a column is added
Now i want the users to be able to add translations, so filling in the empty fields in the translations table if a language has been added. I want to see an overview of all languages with an input field.
Now i have this:
FOR i IN (SELECT * FROM TRANSLATIONS) LOOP
htp.print('<tr><td>' || i.id || '</td>');
FOR a IN (SELECT id, taal, afkorting FROM LANGUAGES) LOOP
l_thevar := 'i.' || a.afkorting;
htp.print('<td><input type=text name=' || i.id || ' value=' || XXXX || '></td>');
END LOOP;
htp.print('</tr>');
END LOOP;
Where the XXXX is, this should be something like i.nl or i.f or i.eg. So what i should be able to do is saying something like take 'i.' and add the nl, f or eg, in my example it's the column a.afkorting. I should paste it together and then plsql should handle it as a variable. It's like a variable variable.
I know it is possible in PHP, but in PL/SQL i haven't been able to find a method to do so.
If anyone could help me, i would really appreciate it!
Tnx a lot!
If something isn't clear, please ask me !

Answer 1

Er.. not sure how to put this gently, but PHP (and Perl/Python/etc) programmers make the absolute WORSE database (PL/SQL) programmers.
The reason is that programmung techniques used in these scripting languages are great.. for scripting languages. But is not just damn ugly in PL/SQL, but wrong, impacting performance, data integrity, violating concurrency principles and so on.
A FOR cursor fetch loop within a FOR cursor fetch loop? Just plain wrong mate. Performance will suck. That code will not scale and will run into a brick wall when the volumes of the underlaying tables grow.
Use a SQL JOIN to join the data sets. And this is an important concept. In PL/SQL (and SQL and in any RDBMS), you must work with DATA SETS - not with rows. This type of row-by-row processing may look similar as processing arrays in PHP. But then a database table is not just a simple array memort struct. So don't treat it like you would an array in PHP.
The correct way to approach this type of problem in PL/SQL is to write a HTML rendering engine that can render any SQL cursor. Does not matter what the cursor contains - the number of columns, data types, etc.
Which means that you can create a SQL SELECT (joining the languages and translation tables) and pass this SQL to the PL/SQL rendering engine to churn into HTML.
Writing a HTML renderer in PL/SQL is not difficult. You use DBMS_SQL cursors. The DBMS_SQL API allows you to create SQL cursors, peek at how many columns are returned, what the column names are, what the data types are - all the data that you need to loop through in order to generate HTML table tags.