Fixed-Width format (.dat, .fixed)

Fixed-width column files are tabular text where every record shares the same column layout: each column begins at a fixed character offset and spans a fixed width, padded with whitespace to fill its slot. Records are separated by line breaks.

The reader splits each line at the configured column boundaries and emits one translatable block per non-empty cell in a column you have marked translatable. Non-translatable cells are preserved as structure so the writer can rebuild the original layout. There is no auto-detection of the layout — you must declare the columns.

Configuration is supplied on the collection's format: block in a .kapi project. The keys are: columns — an ordered list of { name, start, width, translatable } records, where start is the 0-based character offset, width is the column width in characters, and translatable marks the column for extraction (at least one column is required, each with a non-empty name, a positive width, and a non-negative start); hasHeader — when true, treat the first row as a header; and trimValues — when true, trim leading and trailing whitespace from each cell's value.

IDfixedwidth

SourceBuilt-in

Extensions.dat, .fixed

CapabilitiesRead + Write

This format has no configurable parameters.

Examples

Extract a translatable column

A two-column layout where only the second column (the label) is translatable; the first column is an identifier.

format:
  name: fixedwidth
  config:
    columns:
      - { name: id, start: 0, width: 8, translatable: false }
      - { name: label, start: 8, width: 40, translatable: true }
    trimValues: true

Processing notes

• One block per non-empty translatable cell; non-translatable cells are emitted as Data so the layout round-trips.

• With hasHeader enabled the first row is treated as a header rather than data.

• Padding whitespace is preserved by default; set trimValues to strip it from extracted values.

Limitations

• The column layout must be declared explicitly; there is no auto-detection of column boundaries.

• Offsets and widths are counted in characters (runes); multibyte characters occupy one position each, which may differ from byte-oriented layouts.

← Back to the Format Reference