Open richardfoltin opened 6 years ago
The query originally resulted 0 in maxCP if the trainer did not have any pokemon in gyms (also only counted the cp of the pokémon in gyms). I left it this way - assuming this was intended - but it can be fixed easly
First query Current code: 0.0635s Your code: 1.1509s (>1s in first run, second one seems to be cached and is super fast 0.0005s, but this seems to be valid for some seconds and then again takes way longer then the current code)
Second query Current Code: 0.0110s Your code: 0.0047s
Third query Current Code: 0.0087s Your code: 0.0045s
In summary, the current code is always giving me faster responses of the XHR call of aru.php in my setup. Which is due to the first query taking way more time.
~500ms vs. >1s
1.200 gymmember 120.000 gympokemon 10.000 trainer
Is the API live again? :-O
No, but we now can MITM the app and extract proto infos from real devices
Make sure that you have index on the following fileds: trainer.name gympokemon.trainer_name gympokemon.pokemon_uid gymmember.pokemon_uid
Optimized SQL query of loading trainer data
Description
Modifications in the query
Motivation and Context
On bigger databases, trainer page loaded very slowly
How Has This Been Tested?
Tested on my configuration with 32000 rows in trainer table, and 630000 rows in gympokemon table. Query ran about 4-6 times faster
Screenshots (if appropriate):
Types of changes