Data transformation patterns

Importing Data in Java

Anthony Markham

VP Quant Developer

Map function

Transforms each element in a column

DoubleColumn celsius = DoubleColumn.create("Celsius", 0, 10, 20, 30);

DoubleColumn fahrenheit = celsius.map(c -> c * 9.0/5.0 + 32);
table.addColumns(fahrenheit.setName("Fahrenheit"));

Celsius	Fahrenheit
0.0	32.0
10.0	50.0
20.0	68.0
30.0	86.0

Reduce function

Combines all the values in a column into a single result
Applies an accumulator pattern (like sum, max, or custom logic)
Useful for statistical and analytical operations

// Find the total sales
double totalSales = table.doubleColumn("Sales").reduce(0, Double::sum);

// Find the maximum
double largeSales = table.doubleColumn("Sales").reduce(0, (acc, x) -> acc + (x > 5000 ? 1 : 0));

Row iteration with forEach

Iterates through each row in a table

DoubleColumn difference = DoubleColumn.create("Difference");
table.forEach(row -> {
    double celsius = row.getDouble("Celsius");
    double fahrenheit = row.getDouble("Fahrenheit");
    difference.append(fahrenheit - celsius);
});

table.addColumns(difference);

Celsius	Fahrenheit	Difference
0.0	32.0	32.0
10.0	50.0	40.0
20.0	68.0	48.0

Transformation pipelines

Chain multiple operations together
Increase the readability and maintainability of our code ✅
Efficient data processing ✅

Table result = originalTable
    .where(numberColumn("Age").isGreaterThan(18)) // Filter on Age > 18

    .addColumns(
        numberColumn("Income").map(i -> i * 1.1).setName("AdjustedIncome")
    );


// Calculate average income
double avgIncome = result.doubleColumn("AdjustedIncome")
        .reduce(0.0, Double::sum) / result.rowCount();

Recap

map() - apply a function to transform column values
forEach() - iterate through rows to access multiple columns
reduce() - aggregate and summarize data into single values

An image that shows the three data transformation functions

Let's practice!

Importing Data in Java