在与另一列具有相同值的行中选择具有最高列值的行

Question

This time, I will ask about something works generally, but when data is huge it fails.

My case is the same in this post.

how to query for rows that highest column value among rows that have same value for one of the columns

I used

SELECT `ID`, `PDBID`, `Chain`, `UniProtID`, `PDBASequence`, `pI`, `experiment`, `resolution`
FROM protein p 
WHERE `resolution`= (SELECT MAX(`resolution`)
                              FROM protein 
                              GROUP BY `PDBASequence`
                              HAVING `PDBASequence` = p.`PDBASequence`)

I also tried:

 SELECT `ID`, `PDBID`, `Chain`, `UniProtID`, `PDBASequence`, `pI`, `experiment`, `resolution`
    FROM protein p 
    WHERE `resolution`= (SELECT MAX(`resolution`)
                                  FROM protein 
                                  WHERE `PDBASequence` = p.`PDBASequence`)

I have to group by sequences according to PDBASequence. But, at the same time selected representative must be the one which has the max resolution value.

I tried this code on a small set. Working no problem. However, when I tried to run it on real table which has 80980 rows, execution takes almost forever. In addition, my other computer gives Mysql server has gone away error because of the execution type and pocket size. I fixed the settings in my.ini and ran the code again. Nothing changed. Still no result :( What should I do? Thanks I assigned index on resolution in protein table. However,it did not change anything.

Answer 1

See if this version is any better.

SELECT p.`ID`, p.`PDBID`, p.`Chain`, p.`UniProtID`, p.`PDBASequence`, p.`pI`, p.`experiment`, p.`resolution`
FROM protein p 
    INNER JOIN (SELECT PDBASequence, MAX(`resolution`) AS MaxResolution
                    FROM protein 
                    GROUP BY `PDBASequence`) q
        ON p.PDBASequence = q.PDBASequence
            AND p.resolution = q.MaxResolution