Imploding in PHP vs imploding in MySQL - which uses less cpu? -

which of these options more optimal?

imploding in mysql

$rsfriends = $cnn->execute('select cast(group_concat(id_friend) char) friends                                table_friend                               id_user = '.q($_session['id_user'])); $friends = $rsfriends->fields['friends']; echo $friends; 

vs.

imploding in php

$rsfriends = $cnn->execute('select id_friend                                table_friend                               id_user = '.q($_session['id_user'])); while(!$rsfriends->eof) {     $friends[] = $rsfriends->fields['id_friend'];     $rsfriends->movenext(); } echo implode(',',$friends); 

you should know correct ("most optimal") choice going factor of many variables:

1. your database , application hardware.
2. the size of data
3. what load on database , application servers.
4. indexes , other things may affect query execution plan on dataset
5. how actually want use data.
6. network latency between app server , database server

but can analyze program flow arrive @ answer:

imploding in php:

1. execute select query returning friend ids
2. return every single id application.
3. build string in php read them result.

pros:

• flexibility in using raw ids if need them (meaning — if go creating string in mysql, post-process in php, expensive post-process operations negate benefit seen bypassing php in string-compilation stage)
• ids (are?) returned application numbers, not strings (1234567890 10 bytes ascii, 4 bytes 32-bit integer)

imploding in mysql:

1. execute aggregate query
2. build string aggregates results
3. return 1 single string application
4. output string

pros:

• might use less memory on database server large datasets
• can't think of else. need big concatenated string, solution sub-optimal.