is is necessary to learn java for hadoop?

Question

I wanted to start learning the big data technology from the scratch. I wanted to know is it necessary to learn java for operating with hadoop as i am already well versed in python?

Sorry, but this question is off-topic for this site. You can talk about it with your peers. — aIKid, Nov 12 '13 at 14:42
@aIKid how is this question off topic? If you have knowledge of hadoop you can answer it with no problem. — Luiggi Mendoza, Nov 12 '13 at 14:45

score 2 · Answer 1 · answered Nov 12 '13 at 14:44

No, you don't necessarily need java knowledge, as you can write map-reduce jobs perfectly well in pig or hive (similar to SQL). However, as with all layers of abstraction, at some point you may well need to know what is going on "behind the scenes" and being able to look, understand and debug the underlying java is a big advantage.

There is a lot of effort currently going into providing a more complete SQL interface to hadoop, with tools such as Impala (Cloudera), Presto (Facebook), Phoenix and Hive (already mentioned).

Frank Visaggio · Answer 2 · 2013-11-12T15:21:28.657

1

Not needed at all , though thats just my opinion. if you python well you should be fine.

check this out writing a hadoop map reduce in python. theres a lot of ways to implement solutions with hadoop. Just because a great deal of them are in Java doesnt mean java is the only tool to solve use . If your working with legacy that is written in java then knowing the basics may help but to be honest i think you could just reference things as you come across them. There is no need to spend a week learning the intricacies of Java 7 and whats new in Java 8 for your current needs.

edited Nov 12 '13 at 15:21

answered Nov 12 '13 at 14:42

Frank Visaggio

3,642
9
34
71

1

Provide suitable source too. The answer could be more better, opinions don't count as much as credible source (libraries, in this case) – shad0w_wa1k3r Nov 12 '13 at 14:52
1

think i am slightly more on track. thanks for the comment – Frank Visaggio Nov 12 '13 at 15:21
1

Thank you so much. I will stick to python as it is much less complicated language than java. – Abhinav Goel Nov 12 '13 at 23:15

score 1 · Answer 3 · edited May 23 '17 at 12:19

1

Check out MRJob, a python based wrapped for hadoop jobs running, logging and monitoring.

Although pure java solutions might be faster in some cases, you hardly ever will need to debug java code.

edited May 23 '17 at 12:19

Community

1
1

answered Nov 12 '13 at 14:46

alko

46,136
12
94
102

is is necessary to learn java for hadoop?

3 Answers3