Load velocity from Hadoop hive - Qlik Community

Report Inappropriate Content · ‎2014-02-10

Hi all,

We are using ODBC connection by MapR to load tables from hive to QV.

currently a table with 6 columns and 70 million rows takes 35 minutes to Load.

We are going to Load tables with more than 200 columns and more than 50 million rows.

Can somebody advice whether ODBC is the best connection?

regards,

Boris

luis_pimentel · ‎2014-02-10

Do you need to load all data every time? Or just the new/updated rows?

Take a look to this document, It might help you

Best regards,

-Luis.

Report Inappropriate Content · ‎2014-02-10

Thanks,

but in one case the incremental itself is 70 million.

and the other one the whole table is updated every day (also 50+ million).

luis_pimentel · ‎2014-02-10

Ok,

did you try with OLE DB instead of ODBC. As far as I know, OLE DB is always recommended.

Hope that helps.

Report Inappropriate Content · ‎2014-02-10

as far as I know OLE DB doesn't support hadoop hive.

luis_pimentel · ‎2014-02-10

And what about this doc

Report Inappropriate Content · ‎2014-02-10

I understood that JDBC is much slower than the MapR's ODBC.