Field transforms & row rules

A field mapping doesn’t have to copy a source value straight through. Each mapping can apply a transform to change how the value is produced, and a row-selection rule to decide which source rows it reads. Together these let one map handle documents whose data needs a little reshaping before it fits your definition.

Transforms

A transform decides how a source field becomes the mapped value. Each field mapping can use one of:

Direct — use the source value as-is. This is the default.
Date reformatting — read a date in the source’s format and output it in the format your material expects.
Default value — supply a fixed value when the source field is empty or absent.
Concatenation — join several source fields together into one property value.

Row-selection rules

For a repeating source (a CSV or Excel table, or a PDF results table), a row-selection rule controls which rows a field reads:

First match — take the value from the first matching row only.
Filter — restrict to the rows that match a condition, and read from those.
Emit one record per match — produce one output record for every matching row. This is how a repeating entity turns a table into multiple output instances — the same behaviour as turning on Each Row Mode for the entity.

Choosing a row-selection rule for a repeating field

Putting them together

A single mapping can use both. For example: filter a results table to the rows for one test, then emit one record per matching row, and apply date reformatting to a date column so each instance carries a correctly formatted release date.