[wups]

Quote:

Originally Posted by AndreT Simply setting the cache to true in mysql_function_query won't do the trick, you need to use the cache functions in the returning callback as well. cache_get_data returns stores the amount of rows and fields in the given parameters. This is useful for checking validity of what a SELECT query outputs and looping through the results as well. cache_get_row stores the field data in the destination string. It is as simple as that: pawn Код: // Example query: SELECT name,kills FROM players WHERE id = %d LIMIT 0,1 new rows, fields; cache_get_data(rows, fields); if(rows) {     new PlayerName[MAX_PLAYER_NAME], kills, storage[12];     cache_get_row(0, 0, PlayerName);     cache_get_row(0, 1, storage), kills = strval(storage);     printf("Player %s has %d kills!", PlayerName, kills); } I haven't tested cache_get_field, but according to the source, it is supposed to return the name of the given field. There's also cache_get_field_content which will get the data by the field's name, for example: pawn Код: cache_get_field_content(0, "name", PlayerName); cache_get_field_content(0, "kills", storage), kills = strval(storage); I have avoided using this function for the same reason people avoid mysql_fetch_field_row. It loops through the cached array of field names and compares the strings, but in case of cached results, the difference won't of course be as bad as for mysql_fetch_row_format and mysql_fetch_field_row (although the functions are not designed to do the same thing, one is just faster to use when it comes to a large number of fields). I have a house loading/parsing function which runs on startup only. When using the regular method (store the result, put mysql_fetch_row_format in a loop, etc), it took around 2.5 seconds to execute. I converted that function to use the cache method and the differences are slightly astonishing! For the sake of it, I searched them out in my outbox (sent to xxmitsu previously), and for the old method, all mysql_fetch_row calls took ~3000ms, sscanf parsing took ~30ms and other parts of code in the loop ~130ms. After converting, the whole execution time is ~300ms, cache functions take around 160ms of it. I didn't run very many tests, but when starting the server up on Windows, I can clearly tell that the gain in startup is literally seconds!

That's some nice experience.

Anyway, is anyone using the latest version in a LIVE server?