Parse node
RegEx
Match, extract, replace, or split text with a regular expression.
What it does
Applies a regular expression to a text column, four ways: Match adds a true/false column; Replace rewrites matches; Parse pulls each capture group into its own typed column; Tokenize splits matches into columns or explodes them into rows.
Inputs and outputs
| Inputs | One. |
| Outputs | One. |
When to use it
- Pull parts out of messy text (an area code, a domain, an id) into clean columns
- Flag or filter rows whose text matches a pattern
- Standardize or scrub text by rewriting what matches
- Split a repeating or delimited field into columns or rows
Examples
Parse named groups
(?<area>\d{3})-(?<line>\d{4})Makes an 'area' and a 'line' column automatically.
Match
\d{5}(-\d{4})?Adds a true/false column: does it look like a ZIP?
Replace with a group
(\d+)-(\d+) → $2/$1
Swaps the two numbers around.
Tokenize to rows
\w+ → Rows
Each word becomes its own row.
Tips
- Name a group with (?<name>…) and Parse auto-names the column after it.
- Flags go inside the pattern: (?i) case-insensitive, (?m) multiline, (?s) dot matches newlines.
- A value that doesn't match becomes null (Parse), false (Match) or stays unchanged (Replace). It never fails the run.
- Look-ahead/behind and backreferences aren't supported; rewrite a look-around as a capture group and use Parse.
- In Replace, $1 and ${name} insert captured text; write $$ for a literal dollar sign.