I see multiple moving parts here. Using the batch API causes your client to do a single RPC call which is more efficient than multiple. But in my understanding you can also tune your client behaviour by manually flushing your write buffer after multiple put(Put) calls.
I think you need to test your particular scenario to see which one is better. Andor On Fri, May 10, 2019 at 9:48 AM Guillermo Ortiz <[email protected]> wrote: > I thought that but I saw this > > https://stackoverflow.com/questions/28754077/is-hbase-batch-put-putlistput-faster-than-putput-what-is-the-capacity-of > > El jue., 9 may. 2019 19:08, Andor Molnar <[email protected]> > escribió: > > > Hi Guillermo, > > > > I'm not sure which version of HBase you're referring to, but it looks > like > > the API docs are quite clear about these 2 calls: > > "Puts some data in the table." vs "Batch puts the specified data into the > > table." > > > > Essentially you can pass multiple Put commands to the second call > > (listPuts) and it uses the batch API, so it's probably more efficient for > > bulk load scenarios. > > > > Regards, > > Andor > > > > > > > > On Thu, May 9, 2019 at 4:05 PM Guillermo Ortiz <[email protected]> > > wrote: > > > > > I have seen that there are two methods to make a put, I would like to > > know > > > if there's any different between call table.put(put) or > > table.put(listPuts) > > > or under the hood is the same? > > > > > > My use case is to put many records with spark (kind of bulk load) > > > > > >
