Refactoring to Java 8 Trisha Gee (@trisha_gee) Developer & - PowerPoint PPT Presentation

Refactoring to Java 8 Trisha Gee (@trisha_gee) Developer & Technical Advocate, JetBrains

Why Java 8?

It’s Faster • Performance Improvements in Common Data Structures • Fork/Join Speed Improvements • Changes to Support Concurrency • …and more http://bit.ly/refJ8

Easy to Parallelize

Fewer Lines of Code

New Solutions to Problems

Minimizes Errors

Safety Check

Test Coverage

Performance Tests

Decide on the Goals

Limit the Scope

Morphia https:// github.com/mongodb/morphia

Refactoring!

Lambda Expressions

Automatic Refactoring • Predicate • Comparator • Runnable • etc …

Abstract classes

Advanced Search

Collections & Streams API

Automatic Refactoring • For loop to collect • For loop to forEach • …with and without Streams

Manual Refactoring

Optional

But what about performance?

Lambdas

 Anonymous Inner Classes vs Lambdas 180 160 140 120 100 Ops/ms 80 60 40 20 0  Anonymous Inner Class Lambda

Performance of Capture  20 18 16 14 12 nsec/op 10 8 6 4 2 0  single thread max threads anonymous(static) anonymous(non-static) lambda http://www.oracle.com/technetwork/java/jvmls2013kuksen-2014088.pdf

 Logging Performance 500,000 450,000 400,000 350,000 300,000 Ops/ms 250,000 200,000 150,000 100,000 50,000 0 Constant message Variable message  Direct call Lambda

Streams vs Iteration

Iterator vs Stream (1000 elements)  9 8 7 6 5 Ops/ms 4 3 2 1 0  for loop forEach()

IterHelper  7000 6000 5000 4000 Ops/ms 3000 2000 1000 0  1 10 100 1000 original simplified refactored

IterHelper  9 8 7 6 5 Ops/ms 4 3 2 1 0 1000 10000 100000  original simplified refactored

BasicDAO – map & collect

BasicDAO  25000 20000 15000 Ops/ms 10000 5000 0 1 10 100 1000  original simplified refactored

DuplicatedAttributeNames – filter, map & collect

 DuplicatedAttributeNames 1200 1000 800 Ops/ms 600 400 200 0  original refactored

EntityScanner – forEach

 EntityScanner 0.06 0.05 0.04 Ops/ms 0.03 0.02 0.01 0  original refactored

DatastoreImpl – filter & forEach

 DatastoreImpl 0.7 0.6 0.5 0.4 Ops/ms 0.3 0.2 0.1 0 original refactored 

MappingValidator – single stream operation

 MappingValidator 900 800 700 600 Ops/ms 500 400 300 200 100 0 EntityWithOneError EntityWith10Errors EntityWith20Errors original refactored 

QueryImpl – multiple operations

 QueryImpl 3500 3000 2500 2000 Ops/ms 1500 1000 500 0  original simplified refactored

Going parallel

 map() 0.9 0.8 Time Taken (seconds) 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Serial Parallel  MacBook Surface

findAny()  160 140 Time Taken (millis) 120 100 80 60 40 20 0 Serial Parallel  MacBook Surface

Optional

 Compare Constant Field Value with Null 350,000 300,000 250,000 200,000 Ops/ms 150,000 100,000 50,000 0 Value not null Value null  Null check Optional

 Compare Variable Field Value with Null 300,000 250,000 200,000 Ops/ms 150,000 100,000 50,000 0 Value not null Value null  Null check Optional

Summary

Sometimes new idioms decrease clutter

…but sometimes they don’t

Sometimes the new features improve performance

 Anonymous Inner Classes vs Lambdas 180 160 140 120 100 Ops/ms 80 60 40 20 0  Anonymous Inner Class Lambda

…and sometimes they don’t

 DuplicatedAttributeNames 1200 1000 800 Ops/ms 600 400 200 0  original refactored

Sometimes a new feature makes life easier

…sometimes not so much

Some refactoring can safely be done automatically

…often you need a human brain

Conclusion

Should you migrate your code to Java 8?

It Depends

Always remember what your goal is And compare results to it

Understand what may impact performance And if in doubt, measure

Code may magically improve Or you may expose areas for improvement

Your tools can help you But you need to apply your brain too

http://bit.ly/refJ8 @trisha_gee