问题
This is a follow-up to a question I asked previously.
I can calculate the rank (including the logic for ties) just fine; the issue is detecting future duplicates when I come across the first instance of a duplicate rank.
Here is the SQL query to get the result set:
SELECT
s1.team_id,
sum(s1.score>s2.score) wins
FROM scoreboard s1
LEFT JOIN scoreboard s2
ON s1.year=s2.year
AND s1.week=s2.week
AND s1.playoffs=s2.playoffs
AND s1.game_id=s2.game_id
AND s1.location<>s2.location
GROUP BY s1.team_id
ORDER BY wins DESC;
Here is the sample SQL result set which I'll loop through in PHP:
team_id wins
--------------
10 52
2 48
5 46
11 46
3 42
9 39
...
Here is my PHP code for the display, which needs to append "T-" to all tied ranks:
$i = 0;
while($row = mysql_fetch_assoc($r)) { //iterate thru ordered (desc) SQL results
++$i;
($row['wins'] == $prev_val)
? $rnk = 'T-' . $rnk //same as previous score, indicate tie
: $rnk = $i; //not same as previous score
$rnk = str_replace('T-T-','T-',$rnk); //eliminate duplicative tie indicator
if ($row['team_id'] == $team_id) { //current team in resultset matches team in question, set team's rank
$arr_ranks['tp']['cat'] = 'Total Wins';
$arr_ranks['tp']['actual'] = number_format($row['wins'],1);
$arr_ranks['tp']['league_rank'] = $rnk;
$arr_ranks['tp']['div_rank'] = $div_rnk;
}
else if ($i == 1) { //current team is category leader (rank=1) and is not team in question, set current team as leader
$arr_ranks['tp']['leader'] = "<a href='index.php?view=franchise&team_id=" . $row['team_id'] . "'>" . get_team_name($row['team_id']) . '</a> (' . number_format($row['wins']) . ')';
}
$prev_val = $row['wins']; //set current score as previous score for next iteration of loop
}
The "tie" logic above will capture team #4 as having tied with team #3, but not vice versa.
In other words, for team #3, $rnk = 3, while for team #4, $rnk = T-3. (Both should be "T-3".)
So the question becomes: how do I "look ahead" while iterating through the results to find out if the current score is a tie/duplicate of scores further down the list, so I can treat it as a tie along with the subsequent dupes?
@Airzooka gave me a potential solution, but I'm curious to know if there's a more efficient way to do it (possibly at the SQL level even).
Thanks.
回答1:
In pseudo-code:
loop through rows as row1
loop through rows as row2
if row1 ain't row2 and row1.points == row2.points, append T
Update:
Ok, how about this, since you're ordering your result set by wins, anyhow: try storing information about each row in a temporary array or variables, like $previousTeamWins,
$previousTeamName, etc. Then you can compare the current and the previous and assign the T based on that. So you're effectively delaying the assignment until the following iteration (or until the exit of the loop in the case of the final row). One trip through the row set, should get the job done.
回答2:
What about this?
SELECT
t1.*,
EXISTS (
SELECT *
FROM `teams` as t2
WHERE t2.pts = t1.pts AND t1.id != t2.id
) as `tied`
FROM `teams` as t1
...
回答3:
Try this i hope its work for you
SELECT t1.`id`,t1.`pts`,t1.`team_id`,
IF(t1.`pts` = t2.`pts` AND t1.`id` != t2.`id` ,1,0) AS `tied`
FROM `teams` t1
LEFT JOIN `teams` t2 on t2.`pts` = t1.`pts` AND t2.`id` != t1.`id`
GROUP bY t1.`id`
回答4:
Oi, still looking at your code. So at the end of the day, you only want to output one or two teams, right? One if the leader is the team in question, and two otherwise. If that's the case, try something like this (with the warning that it isn't tested):
$i = 0;
while($row = mysql_fetch_assoc($r)) { //iterate thru ordered (desc) SQL results
++$i;
($row['wins'] == $prev_wins)
? $rnk = $prev_rnk //same as previous score, indicate tie
: $rnk = $i; //not same as previous score
$rnk = str_replace('T-T-','T-',$rnk); //eliminate duplicative tie indicator
if ($prev_team_id == $team_id) {
$arr_ranks['tp']['cat'] = 'Total Wins';
$arr_ranks['tp']['actual'] = number_format($prev_wins,1);
$arr_ranks['tp']['league_rank'] = $prev_rnk;
$arr_ranks['tp']['div_rank'] = $div_rnk;
if ($row['wins'] == $prev_wins)
{
$arr_ranks['tp']['tie'] = true;
}
else
{
$arr_ranks['tp']['tie'] = false;
}
break;
}
else if ($i == 1) { //current team is category leader (rank=1) and is not team in question, set current team as leader
$arr_ranks['tp']['leader'] = "<a href='index.php?view=franchise&team_id=" . $row['team_id'] . "'>" . get_team_name($row['team_id']) . '</a> (' . number_format($row['wins']) . ')';
}
$prev_wins = $row['wins'];
$prev_team_id = $row['team_id'];
$prev_rnk = $rnk;
}
if ($prev_team_id == $team_id) {
$arr_ranks['tp']['cat'] = 'Total Wins';
$arr_ranks['tp']['actual'] = number_format($prev_wins,1);
$arr_ranks['tp']['league_rank'] = $prev_rnk;
$arr_ranks['tp']['div_rank'] = $div_rnk;
if ($row['wins'] == $prev_wins)
{
$arr_ranks['tp']['tie'] = true;
}
else
{
$arr_ranks['tp']['tie'] = false;
}
}
回答5:
I think I found a solution via SQL! May not be elegant (I can clean it up later), but here goes...
The query:
SELECT a.team_id, a.wins, count(*) instances
FROM
(SELECT
s1.team_id,
sum(s1.score>s2.score) wins
FROM scoreboard s1
LEFT JOIN scoreboard s2
ON s1.year=s2.year
AND s1.week=s2.week
AND s1.playoffs=s2.playoffs
AND s1.game_id=s2.game_id
AND s1.location<>s2.location
GROUP BY s1.team_id) AS a
LEFT JOIN
(SELECT
sum(s1.score>s2.score) wins
FROM scoreboard s1
LEFT JOIN scoreboard s2
ON s1.year=s2.year
AND s1.week=s2.week
AND s1.playoffs=s2.playoffs
AND s1.game_id=s2.game_id
AND s1.location<>s2.location
GROUP BY s1.team_id) AS b
ON a.wins = b.wins
GROUP BY a.team_id, b.wins
ORDER BY a.wins DESC;
This gives the output...
=================================
|team_id | wins |instances |
=================================
|10 | 44 |1 |
|2 | 42 |3 | //tie
|9 | 42 |3 | //tie
|5 | 42 |3 | //tie
|3 | 41 |1 |
|11 | 40 |1 |
|... | | |
=================================
Then, in PHP, I'll be able to detect all ties by checking when $row['instances'] > 1.
Thanks all for bearing with me through this unnaturally cumbersome issue!
来源：https://stackoverflow.com/questions/11001158/recognizing-duplicates-while-iterating-through-query-results