For maximum throughput you might want to consider IgniteDataStreamer rather than SQL. If you create your cache in Java or XML you could also avoid indexes entirely. There are lots of options here depending on your use case. In any case, I would urge you to measure the performance before (and after) making optimisations.
Cache group isn’t directly exported to the SQL API. You can create a cache template that defines the cache group and use that in your CREATE statement. (https://apacheignite.readme.io/docs/cache-template <https://apacheignite.readme.io/docs/cache-template>) > On 15 May 2020, at 12:52, adipro <aditya.har...@zohocorp.com> wrote: > > I thought it's obvious that, SQL table return in O(log n) with n being number > of elements in BTree. Our task being write intensive, If all the threads are > inserting data to same tree then rebalancing and accessing time would be > affected right? > > So for this sake, if we divide each thread's SQL data as separate. In future > even if the single thread's data increases also, there wouldn't be any > problem right? > > > Actually I didn't know regarding that Cache groups. But I'm getting this > error if I use "GROUPNAME" attribute in CREATE TABLE syntax. Can you please > tell why? > > > > -- > Sent from: http://apache-ignite-users.70518.x6.nabble.com/
