Best thing about it that Its support partition as well as bucketing concepts. And also we can use indexing.The best use of spark sql to implement map reduced to fetch the data the libraries it provide for the map reducer program is very easy to use and write the query.new learner can also learn spark core and spark sql at its beginning if the developer knows the basics sql. Review collected by and hosted on G2.com.

The time it takes for an execution of a query.although we can use indexing but still its takes much time to fetch the results in the database which is not suitable for a higher kind of data analysis or in a learg amount of data we can't prefer spark sql. Review collected by and hosted on G2.com.

What I like most is its in memory processing and its rich API libraries. We can code in Python, Scala, R, Java and most importantly in SQL. It follows lazy evaluation which only triggers when we run actions. It allows us to do cashing, persisting, partitioning, broadcasting and many more which optimise the performance. Review collected by and hosted on G2.com.

Nothing Spark is at its best. I see no drawbacks. Review collected by and hosted on G2.com.

Fast, effecient, scalable, supports python and R now. Mostly in memory vs disks. Supports SparkSQL. Supports variety of libraries not available in Hadoop/MapREduce frameowork Review collected by and hosted on G2.com.

some libs that are not meant to be executed in a distributed, does not work well in Spark. SparkSQL doesn't do ACIDity (ie updates). Limited by memory per node. If not enough, then failure. Review collected by and hosted on G2.com.

I believe spark delivers the good without putting up excessive work. Spark has a wide enough array of features for a communication system to work properly on a business. It is light, easy to use, and simple to understand. Review collected by and hosted on G2.com.

At some point, it might be necessary to update the look and feel of this IM service as its main layout and appearance seems to have fallen behind in modern times. Review collected by and hosted on G2.com.

High level APIs

Multi language support

Clean APIs

Powerful utility methods

Integration with multiple platforms and databases

Fast

Efficient use of Memory

Amazing ML support in pyspark Review collected by and hosted on G2.com.

More support in Java with cleaner methods and more suggestions. Add more about different methods in spark streaming. Review collected by and hosted on G2.com.

The best thing about spark is it provide massive parellel processing on large datatset which made my life so easy in term of time complexity. Review collected by and hosted on G2.com.

What i don't like about spark is complex coding syntax and there is not much documentation around operations. Review collected by and hosted on G2.com.

The best thing about spark is, it uses memory to do all the computation unlike hadoop which uses disk to do mapper and reducer tasks which eventually take a large amount of time. Review collected by and hosted on G2.com.

I feel debugging spark is a bit challenging since out of all mappers one can fail and it's really difficult to view the logs of each mapper to see which one failed Review collected by and hosted on G2.com.

Spark is great for running batch and streaming big data jobs. It is very easy to work with, allowing you to distribute data in a fault tolerant way and allow you to scale out your performance depending on the job size. It's also great to have access to the Python, Scala and R API's. Review collected by and hosted on G2.com.

Join's in Spark can be very finnicky and can often be hard to trace back errors when performing jobs as the stack trace does not hold anything back. Review collected by and hosted on G2.com.

In-memory processing of Spark helps process large amounts of data quickly. Complex analytics and machine learning streams should start leveraging Spark's parallel processing. Review collected by and hosted on G2.com.

There is no support for real-time processing. People who work should have a clear understanding of the data. Sometimes small files have issues processing in a large clustered Spark environment. Review collected by and hosted on G2.com.

Its so simple to communicate with my coworkers Review collected by and hosted on G2.com.

You have to click on contact name to see previous conversations instead of there being a tab with old conversations Review collected by and hosted on G2.com.