Last Segment, One-Based

Last Segment, One-Based

A document-import job hands your extension full file paths straight from the customers' servers: C:\Exports\2026\invoice-1001.pdf from a Windows share one minute, /exports/2026/invoice-1001.pdf from a Linux one the next, and every now and then a path that ends in a separator because someone pointed the job at a folder instead of a file. To fill the attachment's file name and file extension you need the last segment of the path and the extension of that segment — nothing else.

Requirements

Create a codeunit named "Path Helper" with two public procedures:

procedure FileNameOf(Path: Text): Text
procedure ExtensionOf(Path: Text): Text

Rules:

1. FileNameOf returns everything after the last separator in Path. Both / and \ are separators, and one path may mix them. A path with no separator at all is a bare file name — return it unchanged.
2. When Path ends with a separator there is no file name: return an empty text.
3. ExtensionOf returns the part of the file name (what rule 1 returns) after its last dot, without the dot: backup.2026-03-01.tar.gz has the extension gz. Return it exactly as written in the path — no case change, no trimming.
4. A file name without a dot has no extension: return an empty text. A dot in a folder name does not count — /archive.2025/notes has no extension — and neither does a path that ends with a separator.

You can rely on these guarantees about the input: Path is never empty and never consists of separators only; the file name never starts or ends with a dot; there is no whitespace around the segments.

The Text and List types do the work. Text.Split takes any number of separators and returns a List of [Text] with one element per segment — and when the text ends with a separator, its last element is an empty text, which is exactly what rule 2 needs. AL lists are 1-based: the first element sits at index 1 and the last at Count(). List.Get raises a runtime error for any index outside that range, so index 0 is never "the first element" — it is a crash. Text.LastIndexOf follows the same convention: it returns the 1-based position of the last match and 0 when there is no match — never -1 as in most other languages — so a check for a negative result never fires, and treating 0 as a position slices from the start of the text. Text.Substring is 1-based too. The starter is wrong in exactly these two places; each is marked with a // TODO: comment.

Pick object IDs in the range 50100–50199, and reference other objects by name, never by ID.

Examples

What the tests check

The grading tests call both procedures and compare the returned text character for character: a forward-slash path and a backslash path, a path mixing both separators, a trailing / and a trailing \ (both give an empty file name), a bare file name with no separator, an extension after the last of several dots, a file name without a dot, a path whose only dot sits in a folder name, a trailing separator asked for its extension, and an uppercase extension that must come back uppercase. Several file names and extensions are generated at run time, so returning a constant or matching the examples passes nothing. No test passes an empty path.

Learn More

• Text.Split([Text,...]) method — one call, several separators, a List of [Text] back.
• List data type — the instance methods, and the one line that says lists are 1-based.
• Text.LastIndexOf method — the 1-based position of the last match, and what 0 means.
• Text.Substring method — slicing from a 1-based position to the end.