RegexExtract
Pulls a specific piece out of messy text by describing its pattern, such as an invoice number buried in a memo line or digits mixed into a reference field. It returns the first stretch of text that fits the pattern, the fragment itself rather than the whole string.
Inputs
| Input | Accepts | Required |
|---|---|---|
Text | Text | Yes |
Pattern | Text | Optional |
Outputs
| Output | Produces |
|---|---|
Result | Text |
Settings
Typed in on the node or set by the assistant, not wired.
| Setting | Default | Notes |
|---|---|---|
Pattern | — | The pattern to look for, written as a regular expression. \d+ matches the first run of digits, INV-\d+ matches an invoice number like INV-20443, [A-Z]{3} matches a three-letter code. Only the first match is returned. |
Example
Bank statement memos arrive as free text like Payment for INV-20443, net 30.
Set Pattern to INV-\d+ and the node returns INV-20443, a clean invoice
reference you can match against your receivables table. If a memo contains no
invoice number, nothing is returned for that value.
Tips
Patterns are exact about case and spacing. If extracts come back empty, test
the pattern against one real memo first, then widen it: INV[- ]?\d+ also
accepts INV 20443 and INV20443.