This is pretty well implied with Christopher's message, but Drill ships with a Hive storage plugin which puts Hive jars on the default class path. Just as with Drill native UDFs we pick up the default Hive functions in these jars and register them. Another one that was causing some issues was the Hive rand() function, which looks like we are not integrating with properly. For Drill the random() function can be used instead for the correct behavior.
On Tue, Jun 9, 2015 at 5:20 AM, Christopher Matta <[email protected]> wrote: > The function from_unixtime() is actually a Hive UDF and drill doesn’t yet > know how to deal with the 2 extra bytes returned from it. It’s better to > use the Drill function to_timestamp(): > > select to_timestamp(1432912733) from `sys`.`version` > > Chris [email protected] > 215-701-3146 > > On Tue, Jun 9, 2015 at 7:49 AM, Uli Bethke <[email protected]> wrote: > > The following query run through Drill Explorer/ODBC returns NULL, while > > running > > it from SQLLINE it works fine. > > > > select from_unixtime(1432912733) from `sys`.`version`; > > > > Is this a limitation of the ODBC driver? > > Running on Drill 1.0.0 on Windows in Embedded mode. Using 1.0.0 ODBC > driver > > >
