@Mustafa El-badry My recommendation is to downgrade the ojdbc with older version "ojdbc14.jar", then change the ojdbc used in the path where sqoop references (see library paths in $HADOOP_CLASSPATH) and remove the newer "ori18n.jar" to avoid invokation of newer version classes which don't exist in the old ojdbcIf version. This is because you use an old version of Oracle. If this response helped, please vote and accept it as a best answer. I did a bit more research and found that you found the solution yourself. Please post an answer and accept it yourself. You are worth the HCC points. ... View more

@Sergey Paramoshkin Your rdd is getting empty somewhere. The null pointer exception indicates that an aggregation task is attempted against of a null value. Check your data for null where not null should be present and especially on those columns that are subject of aggregation, like a reduce task, for example. In your case, it may be the id field. ... View more

@vamsi valiveti If you look at your statement you will see that group by word and order by word are confused because word is part of the subquery also part of the main query. You need to be explicit on which word you want to order and group by. What was the point to use the alias w, if you are not going to use it? SELECT w.word, count(1) AS count FROM (SELECT explode(split(line,'\\s')) AS word FROM docs) w GROUP BY w.word ORDER BY w.word; ... View more

@Maria Khan What is your Kafka version? Is your topic or other underreplicated? This could be due to a network connection temporary glitch. During the glitch the zookeeper nodes likely couldn't reach each other which leadto topic underreplication issues. You could restart brokers as a workaround. ... View more