Re: What are some use-cases for JOINs?
P.S. I meant to say normalizing rather than de-normalizing. On 21/10/2016 10:36 AM, F21 wrote: Hey all, Normally, rather than de-normalizing my data, I prefer to have the data duplicated in 2 tables. With transactions, it is quite simple to ensure atomic updates to those 2 tables (especially for read-heavy apps). This also makes things easier to query and avoids the memory limits of hash joins. Having said that, what are some use-cases of when JOINs should be used? Cheers, Francis
What are some use-cases for JOINs?
Hey all, Normally, rather than de-normalizing my data, I prefer to have the data duplicated in 2 tables. With transactions, it is quite simple to ensure atomic updates to those 2 tables (especially for read-heavy apps). This also makes things easier to query and avoids the memory limits of hash joins. Having said that, what are some use-cases of when JOINs should be used? Cheers, Francis
Re: Scanning big region parallely
Hi, How to get the Schema out of Hbase using Phoenix driver? metadata.getSchemas() produce.. http://stackoverflow.com/questions/40123547/phoenixioexception-nomethodexception-deferredfileoutputstramstring-string-fi On Thu, Oct 20, 2016 at 11:31 AM, Sanooj Padmakumarwrote: > Hi James, > > We are loading data from Phoenix tables into in-memory database. Based on > the query we are finding the number of phoenix input splits (similar to > what happens inside phoenix MR) and loads the data into in-memory database > in parallel. So we are looking for ways to further parallelize the scan of > a larger region. > > As you mentioned phoenix does this for all its queries. Can you please > provide pointers to the phoenix code where this happens ? > > Thanks for the prompt response. > > Thanks > Sanooj Padmakumar > > On Wed, Oct 19, 2016 at 11:22 PM, James Taylor > wrote: > >> Hi Sanooj, >> I'm not sure what you mean by "loading data in our HBase table into >> in-memory", but Phoenix queries tables in parallel, even within a region >> depending on how you've configured statistics and guideposts as described >> here: http://phoenix.apache.org/update_statistics.html >> >> Thanks, >> James >> >> >> On Wednesday, October 19, 2016, Sanooj Padmakumar >> wrote: >> >>> Hi All >>> >>> >>> We are are loading data in our HBase table into in-memory. For this we >>> provide a start row and end row and scan the hbase regions. Is there a way >>> we can scan a big region in parallel to fasten this whole process ? Any >>> help/pointers on this will be of great help. >>> >>> -- >>> Thanks, >>> Sanooj Padmakumar >>> >> > > > -- > Thanks, > Sanooj Padmakumar >
Re: Scanning big region parallely
Hi James, We are loading data from Phoenix tables into in-memory database. Based on the query we are finding the number of phoenix input splits (similar to what happens inside phoenix MR) and loads the data into in-memory database in parallel. So we are looking for ways to further parallelize the scan of a larger region. As you mentioned phoenix does this for all its queries. Can you please provide pointers to the phoenix code where this happens ? Thanks for the prompt response. Thanks Sanooj Padmakumar On Wed, Oct 19, 2016 at 11:22 PM, James Taylorwrote: > Hi Sanooj, > I'm not sure what you mean by "loading data in our HBase table into > in-memory", but Phoenix queries tables in parallel, even within a region > depending on how you've configured statistics and guideposts as described > here: http://phoenix.apache.org/update_statistics.html > > Thanks, > James > > > On Wednesday, October 19, 2016, Sanooj Padmakumar > wrote: > >> Hi All >> >> >> We are are loading data in our HBase table into in-memory. For this we >> provide a start row and end row and scan the hbase regions. Is there a way >> we can scan a big region in parallel to fasten this whole process ? Any >> help/pointers on this will be of great help. >> >> -- >> Thanks, >> Sanooj Padmakumar >> > -- Thanks, Sanooj Padmakumar