Woohoo! Java 9 has a REPL! Getting Started with JShell and Eclipse January

With Java 9 just around the corner, we explore one of its most exciting new features – the Java 9 REPL (Read-Eval-Print Loop). This REPL is called JShell and it’s a great addition to the Java platform. Here’s why.

With JShell you can easily try out new features and quickly check the behaviour of a section of code. You don’t have to create a long-winded dummy main or JUnit test – simply type away.  To demonstrate the versatility of JShell, I am going to use it in conjunction with the Eclipse January package for data structures. Eclipse January is a set of libraries for handling numerical data in Java, think of it as a ‘numpy for Java’.

Install JShell

JShell is part of Java 9, currently available in an Early Access version from Oracle and other sources. Download and install Java 9 JDK from http://jdk.java.net/9/ or, if you have it available on your platform, you can install with your package manager (e.g. sudo apt-get install openjdk-9-jdk).

Start a terminal and run JShell:capture1

As you can see, JShell allows you to type normal Java statements, leave off semi-colons, run expressions, access expressions from previous outputs, and achieve many other short-cuts. (You can exit JShell with Ctrl-D.)

Using JShell with Eclipse January

To use Eclipse January, you need to:

1. Download January:

Get the January 2.0.2 jar ( or older version January 2.0.1 jar).

2. Download the dependency jars:

The January dependencies are available from Eclipse Orbit, they are:

3. Run JShell again, but add to the classpath all the jars you downloaded (remember to be the in the directory you downloaded the jars to):


"c:\Program Files\Java\jdk-9\bin\jshell.exe"  --class-path org.eclipse.january_2.0.2.v201706051401.jar;org.apache.commons.lang_2.6.0.v201404270220.jar;org.apache.commons.math3_3.5.0.v20160301-1110.jar;org.slf4j.api_1.7.10.v20170428-1633.jar;org.slf4j.binding.nop_1.7.10.v20160301-1109.jar


jshell --class-path org.eclipse.january_2.0.2.v201706051401.jar:org.apache.commons.lang_2.6.0.v201404270220.jar:org.apache.commons.math3_3.5.0.v20160301-1110.jar:org.slf4j.api_1.7.10.v20170428-1633.jar:org.slf4j.binding.nop_1.7.10.v20160301-1109.jar

Some notes:
Some version of jshell the command line argument is called -classpath instead of --class-path
If you are using git bash as your shell on Windows, add winpty before calling jshell and use colons to separate the path elements.


Then you can run through the different types of January commands. Note JShell supports completions using the ‘Tab’ key. Also use /! to rerun the last command.

Import classes

Start by importing the needed classes:

import org.eclipse.january.dataset.*

(No need for semi-colons and you can use the normally ill-advised * import)

Array Creation

Eclipse January supports straightforward creation of arrays. Let’s say we want to create a 2-dimensional array with the following data:

[1.0, 2.0, 3.0,
 4.0, 5.0, 6.0,
 7.0, 8.0, 9.0]

First we can create a new dataset:

Dataset dataset = DatasetFactory.createFromObject(new double[] { 1, 2, 3, 4, 5, 6, 7, 8, 9 })

This gives us a 1-dimensional array with 9 elements, as shown below:

[1.0000000, 2.0000000, 3.0000000, 4.0000000, 5.0000000, 6.0000000, 7.0000000, 8.0000000, 9.0000000]

We then need to reshape it to be a 3×3 array:

dataset = dataset.reshape(3, 3)

The reshaped dataset:

 [[1.0000000, 2.0000000, 3.0000000],
 [4.0000000, 5.0000000, 6.0000000],
 [7.0000000, 8.0000000, 9.0000000]]

Or we can do it all in just one step:

Dataset another = DatasetFactory.createFromObject(new double[] { 1, 1, 2, 3, 5, 8, 13, 21, 34 }).reshape(3, 3)

Another dataset:

 [[1.0000000, 1.0000000, 2.0000000],
 [3.0000000, 5.0000000, 8.0000000],
 [13.000000, 21.000000, 34.000000]]

There are methods for obtaining the shape and number of dimensions of datasets


Which gives us:

jshell> dataset.getShape()
$8 ==> int[2] { 3, 3 }

jshell> dataset.getRank()
$9 ==> 2

Datasets also provide functionality for ranges and a random function that all allow easy creation of arrays:

Dataset dataset = DatasetFactory.createRange(15, Dataset.INT32).reshape(3, 5)

[[0, 1, 2, 3, 4],
 [5, 6, 7, 8, 9],
 [10, 11, 12, 13, 14]]

import org.eclipse.january.dataset.Random //specify Random class (see this is why star imports are normally bad)
Dataset another = Random.rand(new int[]{3,5})

[[0.27243843, 0.69695728, 0.20951172, 0.13238926, 0.82180144],
 [0.56326222, 0.94307839, 0.43225034, 0.69251040, 0.22602319],
 [0.79244049, 0.15865358, 0.64611131, 0.71647195, 0.043613393]]

Array Operations

The org.eclipse.january.dataset.Maths provides rich functionality for operating on the Dataset classes. For instance, here’s how you could add 2 Dataset arrays:

Dataset add = Maths.add(dataset, another)

Or you could do it as an inplace addition. The example below creates a new 3×3 array and then adds 100 to each element of the array.

Dataset inplace = DatasetFactory.createFromObject(new double[] { 1, 2, 3, 4, 5, 6, 7, 8, 9 }).reshape(3, 3)

[[101.0000000, 102.0000000, 103.0000000],
 [104.0000000, 105.0000000, 106.0000000],
 [107.0000000, 108.0000000, 109.0000000]]


Datasets simplify extracting portions of the data, known as ‘slices’. For instance, given the array below, let’s say we want to extract the data 2, 3, 5 and 6.

[1, 2, 3,
 4, 5, 6,
 7, 8, 9]

This data resides in the first and second rows and the second and third columns. For slicing, indices for rows and columns are zero-based. A basic slice consists of a start and stop index, where the start index is inclusive and the stop index is exclusive. An optional increment may also be specified. So this example would be expressed as:

Dataset dataset = DatasetFactory.createFromObject(new double[] { 1, 2, 3, 4, 5, 6, 7, 8, 9 }).reshape(3, 3)
Dataset slice = dataset.getSlice(new Slice(0, 2), new Slice(1, 3))

slice of dataset:

[[2.0000000, 3.0000000],
 [5.0000000, 6.0000000]]

Slicing and array manipulation functionality is particularly valuable when dealing with 3-dimensional or n-dimensional data.


For more on Eclipse January see the following examples and give them a go in JShell:

  • NumPy Examples shows how common NumPy constructs map to Eclipse Datasets.
  • Slicing Examples demonstrates slicing, including how to slice a small amount of data out of a dataset too large to fit in memory all at once.
  • Error Examples demonstrates applying an error to datasets.
  • Iteration Examples demonstrates a few ways to iterate through your datasets.
  • Lazy Examples demonstrates how to use datasets which are not entirely loaded in memory.

Eclipse January is a ‘numpy for Java’ but until now users have not really been able to play around with it in the same way you would numpy in Python.

JShell provides a great way to test drive libraries like Eclipse January. There are a couple of features that would be nice-to-have such as a magic variable for the last result (maybe $_ or $!) and maybe a shorter way to print a result (maybe /p :-). But overall, it is great to have and finally gives Java the REPL and ability to be used interactively that many have gotten so used to with other programming languages.

In fact we will be making good use of JShell for the Eclipse January workshop being held at EclipseCon France, see details and register here:  https://www.eclipsecon.org/france2017/session/eclipse-january


Technical Debt: How Do You Unfork a Fork?


Everyone knows how to fork that interesting open source project, it’s simple and handy to do. What’s not so easy to do is to merge back a fork that has over the years taken on a life of its own and for many reasons has diverged drastically from the original repo.

This is a case study of an ongoing project we are doing with SWT XYGraph, a visualisation project that is now part of Eclipse Nebula. It is the story of a fork of SWT XYGraph maintained by Diamond Light Source, the UK’s national synchrotron. But mostly it is a story about the efforts to merge the fork, reduce technical debt, and work towards the goal of sharing software components for Science, a key goal of the Eclipse Science Working Group. Continue reading “Technical Debt: How Do You Unfork a Fork?”

Why You Should Send Flowers To Open Source Developers


You are probably thinking flowers are dumb: they die and they don’t do anything; they are for dates, weddings and funerals. Besides, open source developers do not need flowers, they need money. But nobody wants to pay for software, let alone open source software. Even though today open source software is as vital as roads and bridges, nobody wants to pay to maintain it.

So by all means let’s figure out a way to get more money for open source developers. But in the meantime, here’s why you should still send flowers.

  1. Flowers show appreciation

Taskwarrior share some great lessons learnt for open source maintainers. This includes cautionary tales like: “People will pick a fight with you about all your incidental choices. Your issue tracker, your branching strategy, your version numbers, the text editor you use, and so on.” Open source developers are a tough bunch, but are still likely to suffer from burnout. Sending flowers is the ultimate gesture of positivity and understanding. Plus there is all sorts of research out there that suggests flowers & plants improve mood and promote creativity in the workplace.

  1. People think you are more capable & emotionally intelligent

If altruistic reasons are not enough, then consider this. The act of giving flowers has benefits for the giver. The act of giving makes people happy. There is also research that says people who send flower are perceived as having higher emotional intelligence, capable and courageous.

  1. You will be a better leader

Small things can make all the difference. This idea is partly inspired by my favourite blogger, who wrote about the effect receiving flowers from Marc Benioff, CEO of Salesforce had on her. Basically really good leaders are always thinking of thoughtful ways to get the best out of everybody.

OK so now I have a confession to make. Maybe I am telling you this because Kichwa Coders have recently joined the realm of open source maintainers, heading up and contributing to the January project  (numpy & datastructures for Java). And we are gearing up, together with the rest of the Eclipse Science Working Group, for our first major science release. It is an amazing project that will provide tool infrastructure for countless scientific projects. So it will be awesome, but there is still some trepidation when you cross the threshold from open source user to open source maintainer. It is a little bit like how when you have your own kids, you finally really appreciate your parents and wonder how they did it. Right now I am in awe of open source leads and maintainers everywhere. So just do it. Send flowers to open source developers. And while you’re at it, send some to your parents too.

“Hello World!” – One Small Step for Python Scripting in Eclipse


We are excited to be able to run a ‘hello world’ command using EASE and CPython in Eclipse (quickly followed by running the ‘Zen of Python’ for good measure :). It is the first key stage of proving our approach of using Py4J as the enabling technology to getting scripting available with EASE for Eclipse. It is a small step, but a significant leap towards unprecedented automation, dramatically easier Eclipse extensions and powerful third-party library integration using Python. Continue reading ““Hello World!” – One Small Step for Python Scripting in Eclipse”

Kichwa Coders Join Eclipse Science Working Group Steering Committee

Kichwa Coders is pleased to announce that we have become a member of the Steering Committee of the Eclipse Science Working Group. Kichwa has been an active member of the Science Working Group from the beginning and its involvement includes:

  • Leading projects related to Python scripting in Eclipse EASE
  • Organizing Eclipse community meetups in London (join us for our next one!)
  • Jonah Graham is project co-lead for the January Project for common data structures.
  • Tracy Miranda is on the Program Committee promoting the Science track for Eclipsecon France

Tracy Miranda will represent Kichwa Coders as the Steering Committee participant. The current Steering Committee is made up of members from IBM, Diamond Light Source, Oakridge National Labs and Itema. Kichwa Coders look forward to working closely with other Steering Committee members to shape the future direction of the group and encourage other organizations to participate in this vibrant ecosystem where advancing open-source software advances science.