Dotwave Combining columns
One of the most common feature-engineering moves is taking two columns you already have and combining them into a single new one. Sometimes that means doing maths on two numeric fields; sometimes it means stitching two text fields together. Dotwave handles both in Step 4, Feature Engineering, and in each case the result is a clean new named column added to your dataset without touching the originals. This article walks through arithmetic combining and text concatenation, then shows how the result and its audit record appear.
Arithmetic combine
Arithmetic combining takes two numeric columns and applies an operation between them, referencing each column by name. The four operations available are add, subtract, multiply, and divide. Typical uses include:
total = price + taxto add two amounts.profit = revenue − costto subtract one column from another.line_total = quantity × unit_priceto multiply.rate = wins ÷ attemptsto divide one column by another.
You name the two source columns and the operation, and Dotwave computes the result row by row into a new column. The original columns are left exactly as they were — combining is additive, never destructive.
When you divide two columns, Dotwave checks the denominator column for zeros. If any denominator value is zero, it warns you before the column is created so you can decide how to handle those rows.
Concatenate text columns
Concatenation joins text columns into a single combined string column. The classic example is building a full name from parts — first_name + " " + last_name producing a full_name column — but the same mechanism assembles addresses, composite identifiers, or any label that spans several fields. You choose the text columns to join and any separator between them, such as a space, comma, or hyphen, and Dotwave writes the joined result into one new column while leaving the source fields intact.
How the result appears
Whichever method you use, the outcome is the same shape: a new, named column appended to your dataset and available to every downstream step, chart, and export. You give it a name when you create it, so it reads clearly in the data grid and in your exported file. And because combining is a feature-engineering operation, it is recorded in the audit trail as an added field — the log captures that the column was derived by combining specific source columns, giving anyone reviewing the dataset a clear record of where the field came from rather than an unexplained new column.
For text columns, concatenate creates a single combined column. For numeric columns, use arithmetic operations.
Because the source columns are preserved, you can combine the same fields more than once in different ways — an add for a total and a divide for a rate, both drawn from the same two columns. Each combination is an independent step in your recipe, so your final dataset stays transparent about exactly how every derived column was built.