多个MySQL查询还是一个具有多个JOIN的查询？

Question

Normally the obvious answer to me would be "several joins." But here's the situation: I have a table with a couple thousand entries, tableA. I have another table with several thousand entries, tableB. Each entry has a number the correlates with the id of a row in tableA. There will be multiple rows in tableB that go with a single row in tableA. There will be a third table, tableC, with tens of thousands of entries. It will have multiple rows that correspond to a single row in tableB.

The query might look something like this: (the specifics don't really matter)

SELECT tableA.*, tableB.*, tableC.* LEFT JOIN tableB ON tableB.idA = tableA.id LEFT JOIN tableC ON tableC.idB = tableB.id WHERE tableA.id = some_number;

This works just fine and dandy, but I get tons of duplicate rows because of the joining. I want ONE result for tableA, a few results from tableB, and several from tableC.

Unfortunately to do this I use many nested PHP arrays and for loops. It gets to be messy, so I'm wondering if I should just use 3 separate queries to get rid of all the nested PHP loops, or if I should stick with the JOINS?

Additional info: "tableA": jobs id (primary key) name description date hours customer status

"tableB": work id (primary key) job (correlates with id of row in jobs) summary description date

"tableC": times id (primary key) work (correlates with id of row in work) start end

Desired output: I'm not really sure what kind of format would be best. I want one job row, with a list of work that goes with it, and a list of times that goes with each work.

Example of nested loops: $work is an array with two values. The first is an array of works, the second is an array of times.

foreach ($work[0] as $workk=>$workv) {
    echo("<div><h1>" . $workk . ": " . $workv["summary"] . "</h1><br/>Job: " . $workv["job"] . "</br>Date: " . $workv["date"] . "<br>Description:" . $workv["description"] . "<br/>Times:<br/>");
    foreach ($work[1] as $workid=>$times) {
        if ($workid == $workk) {
            foreach ($times as $time) {
                echo("Start: " . $time[0] . " | End: " . $time[1]);
            }
            break;
        }
    }
}

I get the "work" array using this code in a function (where $query->result() is a row from the long JOIN query)

foreach ($query->result() as $row) {
                if (!array_key_exists($row->id, $work)) {
                    $work[$row->id] = array("job"=>$row->job, "summary"=>$row->summary, "description"=>$row->description, "date"=>$row->date);
                }
                $times[$row->id][] = array($row->start, $row->end);
            }
            return array($work, $times);

Answer 1

You have two* ways of querying the DB:

1. Get all the data together with two left joins.
2. Get the master data and load the details on demand.

* Do not load all the three tables separately and then try to link them in php

The first approach may take a bit longer to load at first, but as the data is loaded there will be no more DB calls. On the other hand, although it might be easy to show the data as a table (with repeating masters), often that is not the desired UI. Unless you are using a special UI component to automatically handle the master-details, you may have some difficulty converting the table to a more sensible form (in this case your loops).

Your problem is in fact a UI design problem as how to show two levels of master-details. There are many UI patterns to show master-details.

For example, you can only load the first level (table A) in a drop down and query the DB and show the second level on drop down change. Or you could have a tree like structure that queries the db on collapse.

Take look at this link , to get some idea. You could also try and search for master-detail UI components for php.