Java

Java came out of Sun Microsystems in 1995 with the promise “write once, run anywhere”: it compiles to bytecode for the Java virtual machine, not for one kind of processor. It’s one of the most-used languages in the world, mostly in business systems and Android apps. Data science reaches it through the big-data tools built on the JVM, such as Hadoop, Spark, and Kafka.

Its standard library is large, with collections, streams, and file and date handling, but it has no statistics or matrices, so the methods here are written out by hand. The implementations use no build tool and no third-party library: each is one file, compiled with javac and run with java. Libraries such as Apache Commons Math aren’t installed.

Quirks

• == compares objects, not their contents. new String("rosa") == "rosa" is false; .equals("rosa") is true.
• The same goes for boxed numbers, with a twist: two Integers holding 127 are ==, but two holding 1000 aren’t, because Java reuses the small ones.
• printf follows the computer’s language settings, so on a German system %.6f prints 29,677500. Every program here passes Locale.ROOT.
• A method that can throw a checked exception, such as IOException when reading a file, must say throws IOException or catch it, or the file won’t compile.
• int arithmetic wraps without warning: Integer.MAX_VALUE + 1 is −2147483648. -7 / 2 is −3; Math.floorDiv(-7, 2) is −4.

git clone https://github.com/qnixon/dswithx.git
cd dswithx
python3 harness/run_tests.py --language java