I was just saying that you don'y need the %pyspark, just override the hive connection string and use it. We did that for Oracle. Eran
בתאריך שבת, 1 באוג׳ 2015, 00:33 מאת Dhaval Patel <dhaval1...@gmail.com>: > Hi Eran, > > That is what I am trying to figure out - how can I put jar in the class > path? I tried adding export var in setup script as mentioned below and it > doesn't seem to add in the classpath. Or did I misunderstood anything? > > Thanks, > Dhaval > > On Fri, Jul 31, 2015 at 5:14 PM, IT CTO <goi....@gmail.com> wrote: > >> As a simple hack you can put the jar in the class path and tgen set the >> jdbc parameters in the hive interpreter parameters. Then use %hive and just >> write sql aginst teradata. >> Eran >> >> בתאריך שבת, 1 באוג׳ 2015, 00:08 מאת Dhaval Patel <dhaval1...@gmail.com>: >> >>> Hi, >>> >>> I am trying to connect to Teradata from spark and getting below error >>> for not finding suitable drivers. >>> >>> : java.sql.SQLException: No suitable driver found for >>> jdbc:teradata://XXXXXX >>> >>> >>> I have tried adding jar files using %dep, as well as in zeppelin-env.sh >>> setting up SPARK_CLASSPATH variable but instead of adding under >>> classpaths, it adds under spark.driver.extraClassPath. >>> SPARK_CLASSPATH=/...path/terajdbc4.jar:/..path/tdgssconfig.jar >>> >>> >>> Below is code I tried from Z : >>> >>> %pyspark >>> df = sqlContext.load(source="jdbc", url="jdbc:teradata://XXXXX, >>> user=XXXXXX, password=XXXXXXX", dbtable="XXXXXXX") >>> >>> I have tried from shell adding the driver and connecting from there and >>> it worked like charm. >>> >>> Thanks in advance! >>> >>> -Dhaval >>> >> >
