I cleared the certification with 9 out of 10 questions , 1 with wrong output as per their expectation.
I have written the exam on 2nd Jan 2016, they will take 3-5 days to validate the exam.
I will share my experience with you & the pattern, you can follow it.
1. Every question in hands-on, no objective
2. Every problem we need write the code & execute on their cdh5 cluster must.
3. Result must & should be their expected form, if not 0 score
examples:
- expected output x,y => you got output x, y or x , y
- they will not consider as a correct answer also.
- such a minor points also you have to remember.
- In code also very carefully we need to write it. other wise they will make it as a wrong answer only.
- simple example for that code need to written in single line in spark scala or python code but readability purpose you might return in multiple lines then the answer will treat as a wrong
- val wordcount = file.flatMap(line => line.split(" ").map( word => (word, 1))
- We can above statement like this also.
- val words = file.flatMap(line => line.split(" ")
- val wordcount = words.map( word => (word, 1))
4. Please read the questions clearly .. because
- some questions they will give partial solution file in specific location on cdh5 vm , we need to give the correct solution on that file only
- some questions we need to read data from hdfs
- some questions we need to read data from local file system
- some questions we need to read data from RDBMS
- some questions we need to read data from HDFS in different file formats also using spark with scala or python or hive or sqoop
- Like avro, parquet, json
- few questions they given partial information only.
- I got 1 question like that.
- I almost waste my time 10 mins there only. finally solved it by deeper observation in data.
5. I faced some issues in writing exam
- font is very low, difficult to read the questions & giving the answers also
- cluster is very slow, don't do multiple tasks it hangs
- same issue i faced, almost it is taken for me 2-3 mins to come up again. like 2 -3 times happen also.
- it may kill your time around 5-10 mins also
- ***Remember locations of normal cdh5 software's and this exam clustercdh5 will be differ as i observed.
- Because of this 1 Question is very difficult to answer from my side. taken more time to solve.
- I used my previous experience on Big Data knowledge to solve that problem. Other wise i may lose the 1 Question answer.
6. Hope you got an idea. What the steps to observe in exam.
7. I am happy to share my experience with this Exam.
8. If any one looking Spark & Scala Training in Hyderabad, you can reach me also.
9. Follow me on below links.
10. check my youtube videos also, going with my channel is "kalyan hadoop"
11. Thanks for your time.
No comments:
Post a Comment