Jandas
Java Data Frames
Bringing the Data Frame concept to Java. It aims to provide features that will be familiar those who have used data frames in Pandas and R.
Main Features
- Mixed type storage for optimal memory usage
- Pretty printing
- Column operations
- Reading / Writing to CSV
- Reading / Writing using SQL
- Group by
- Aggregation
- Multi-level indexing
- Join and merge
- Linear algebra via EJML
Installation
The source code is hosted on GitHub at: https://github.com/PrimePotato/jandas.
It is also available via Maven Central and so the pom.xml should include:
<dependency>
<groupId>io.github.primepotato</groupId>
<artifactId>jandas</artifactId>
<version>0.1</version>
</dependency>
Usage
Manual Construction
Here is a simple example of a data frame being manually constructed:
String[] str1 = {"a", "b", "c"};
StringColumn sc1 = new StringColumn("str", true, str1);
int[] data1 = {1, 2, 3};
IntegerColumn ic1 = new IntegerColumn("int1", true, data1);
int[] data2 = {1, 2, 3};
IntegerColumn ic2 = new IntegerColumn("int2", true, data2);
double[] dblData = {1., 2., 3.};
DoubleColumn dbl1 = new DoubleColumn("dbl1", true, dblData);
List<Column> cols = new ArrayList<>();
cols.add(sc1);
cols.add(ic1);
cols.add(ic2);
cols.add(dbl1);
df1 = new DataFrame("1", cols);
Read & Write to CSV
It is possible to load data from CSV and also save data frame in
dataFrame = Jandas.readCsv("src/test/resources/biostats.csv");
Write to csv:
dataFrame.toCsv("example.csv");
Printing
Here is an example using the sample csv data
dataFrame = Jandas.readCsv("src/test/resources/biostats.csv");
dataFrame.print();
output:
Sex | Weight (Sep) | Weight (Apr) | BMI (Sep) | BMI (Apr) |
---------------------------------------------------------------------
M | 72 | 59 | 22.02 | 18.14 |
M | 97 | 86 | 19.7 | 17.44 |
M | 74 | 69 | 24.09 | 22.43 |
M | 93 | 88 | 26.97 | 25.57 |
F | 68 | 64 | 21.51 | 20.1 |
M | 59 | 55 | 18.69 | 17.4 |
F | 64 | 60 | 24.24 | 22.88 |
F | 56 | 53 | 21.23 | 20.23 |
F | 70 | 68 | 30.26 | 29.24 |
F | 58 | 56 | 21.88 | 21.02 |
... | ... | ... | ... | ... |
M | 66 | 71 | 21.4 | 22.97 |
F | 52 | 57 | 17.72 | 19.42 |
M | 71 | 77 | 22.26 | 23.87 |
F | 55 | 60 | 21.64 | 23.81 |
M | 65 | 71 | 22.51 | 24.45 |
M | 75 | 82 | 23.69 | 25.8 |
F | 42 | 49 | 15.08 | 17.74 |
M | 74 | 82 | 22.64 | 25.33 |
M | 94 | 105 | 36.57 | 40.86 |
Column Operations
Operators such as addition, multiplication, etc are supported on floating point columns. The full details can be found here EJML.
Basic Operation
It is possible to perform operations on columns via the EJML library. For example
dfFreshmen = Jandas.readCsv("src/test/resources/freshman_kgs.csv");
DoubleColumn dc = dfFreshmen.column("BMI (Sep)")
System.out.println(ask.scale(2));
Equations
Also leveraging the EJML library it is possible to create columns from equations:
DoubleColumn a = dfFreshmen.column("BMI (Sep)");
DoubleColumn b = dfFreshmen.column("BMI (Apr)");
DoubleColumn c = new DoubleColumn("Mid", false, new double[0]);
Equation eq = new Equation();
eq.alias(a.getMatrix(), "a", b.getMatrix(), "b", c.getMatrix(), "c");
eq.process("c = (a+b)/2");
Note: that this currently only works on columns with a double data type
Or more simply:
dfFreshmen.createColumn("avg", "(BMISep+BMIApr)/2.0");
Grouping & Aggregation
m
Contributing
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change. Full guidance can be found in Contribution guidelines for this project.