Group by php

[hash]

Well, if you meant $var = 'html' or return 'html', sure. Other than that there is no diff. I just find syntax highlighting to be useful

[paul_wilkins]

That's a good point. Are there any good standardised OOP techniques that allow for proper highlighting of HTML code?

[hash]

Any particular examples? Keep it short I guess, ie make methods that deal with form elements, rather than the whole form.

[paul_wilkins]

Sorry, I'll try to be more clear.

I'm wondering if there are any PHP coding techniques that allow us to use HTML code in such a way that it is not output immediately to the screen (thus satisfying OOP techniques) while also allowing the HTML code to be syntax highlighted by our code editors (thus satisfying our sense of readability).

Edit: s/out/our

[ma201dq]

Use the count query to get table two above, use php to create links for each row, pass the price in the link, use a where query to get all rows for that price. Is that what you mean?

Yeah that’s what I meant, I actually thought there was a neater way of doing it, but thanks everyone for the input.

[ma201dq]

SELECT
     p.price
     ,p.prices
     ,t.id
  FROM
     (SELECT
           t.price
           ,COUNT(*) prices
        FROM
           my_table t
       GROUP
          BY
           t.price) p
 INNER
  JOIN
     my_table t
    ON
     p.price = t.price

The other solution is to eliminate the subquery and count the number of number of rows per price in PHP. In either case your going to need to format the result set on the application side. So I would run a simple select for everything then handle the aggregate calculations on the application side while parsing the result set into a price => rows pair. You could also do as hash said and run the count query then a query for each individual row that pertains to the given price.

this is well thought too, more towards what I was looking for

[Shrapnel_N5]

You have to think of interface first.
First figure out what html you want and what happens if user click a link.
After this you can ask for the SQL query.

[Shrapnel_N5]

but only when you're not planning to store the html code for later use.

Hey Paul.
But why we need that?
In the proper planned application we have all HTML in one place - in the template. And only use for it is output.
And template is more HTML than any other language - another reason to use third method.

[paul_wilkins]

Yep, links need to jus retrieve data fro the database (be immutable), because spiders will be crawling all over them and people can bookmark them and perform other crazy stuff with them.

Buttons are for actions that cause a data change to the server, whether they create, update, or delete information in the database.

Yeah, *hangs head* - I was using CRUD there.

[paul_wilkins]

Hey Paul.
But why we need that?
In the proper planned application we have all HTML in one place - in the template. And only use for it is output.

Because I don't know about you, but many times we are working with other people's code, not to mention house-standards that may not be agreeable to be followed, but must be none-the-less.

[ma201dq]

You have to think of interface first.
First figure out what html you want and what happens if user click a link.
After this you can ask for the SQL query.

yeah I usually do adapt a prototype approach when web deving, actually always. I dont know, but with web dev I hardly ever approch it from data structure point of view, unless I'm deving a web application. what is ur approach?

[Shrapnel_N5]

Unlike desktop application, web application has discrete nature.
Actually, each page acts more like separate application than part of main program.

Therefore, if you have a link on your page, which leads to another page, there are two separate programs to run. And second one don't have access to first one's data and you have to fetch it again. So, no need to fetch actual prices on the first page and there is no use for oddz's query.

On the other hand, if you plan to fill some javascript array with this data and use some DHTML action instead of a hyperlink, you will need all your data. So, there is use for oddz'q query.

That's why you have to go for interface first.

I apologize for my crude English. Such explanations are quite complicated for me.

[SpacePhoenix]

Never rely on javascript being enabled on a given computer. Any web page or web app should be able to function (albeit slightly slower) with javascript disabled or blocked.

[paul_wilkins]

Hear hear.

There is an on-going discussion/argument/dollybrook going on at the moment in the JavaScript forum about How to build a desktop like application interface using JavaScript? where the common consensus is to start with a base HTML page and use CSS for presentation. Get the application working in a non-scripted environment, and only then use JavaScript to improve the user interface/experience.