OCR Extract Pattern

OCR Extract Pattern

Type: ocr-extract-pattern  ·  Package: OCR Activities v1.2.5  ·  Output: Dictionary<string, string>

Since: v1.2.0

Combines OCR + regular expressions in one step: it recognizes the text of an image and, on that text, applies a set of patterns (regex) to extract specific fields. It returns a dictionary { field: value }. It is the most direct way to get structured data out of a scanned document (invoice number, date, total).

How it works

Runs OCR on the image (once) and, on the resulting text, evaluates each pattern of the patterns dictionary:

• It takes the first match of each pattern in the whole text.
• If the pattern has a capture group (...), the value is what the first group captures; otherwise, the whole match.
• If the pattern does not match (or is empty), the field gets empty text "".
• Patterns are case-sensitive; to ignore case, start the pattern with (?i).

Required parameters

Parameter	Type	Default	Description
path	string	—	Path to the image file (.png, .jpg, .bmp, .tiff).
patterns	Dictionary	—	Dictionary fieldName → regex pattern. Use a capture group to extract only the value. Values can be expressions.

Optional parameters

Parameter	Type	Default	Description
language	string	"eng"	Text language: Tesseract code (eng, spa, eng+spa) or regional tag (es-CO). See language.
preprocess	string	"none"	Prior cleanup: none · auto · deskew · binarize. See preprocessing.

Output

Returns a Dictionary<string, string> with one key per pattern in patterns (always all of them) and its text value: the data found, or "" if there was no match. Check for empty fields before using them.

Example

Extract the number, date and total of a scanned invoice:

OCR Extract Pattern
  path     = invoicePath
  language = "eng"
  preprocess = "auto"
  patterns = {
     "number": "INV-(\d+)",
     "date":   "Date:\s*(\d{4}-\d{2}-\d{2})",
     "total":  "Total:\s*\$?([\d.,]+)"
  }
  → output: data

If   condition = isEmpty(data["total"])
 └─ then:
      Throw   message = "Could not read the invoice total"
Log   message = "Invoice " + data["number"] + " — Total " + data["total"]

One capture group per field

The parentheses (...) in the pattern mark which part to extract. INV-(\d+) on INV-021 returns 021. Learn the syntax in Expressions › Regex.

Common errors

Message	Cause and fix
'patterns' is required.	patterns is missing.
'patterns' must be a Dictionary<string, string> mapping field names to regex patterns.	patterns is not a { field: pattern } object (for example, a list or a text).
Invalid regex for field '…': …	That field’s pattern has a syntax error (unclosed parenthesis, unescaped \…).
File not found.	The path does not exist on the robot’s machine.
Tesseract language '…' not found at '…'. Available: [eng, spa].	The language code is not installed.

Related activities

• OCR Recognize File — get the text and extract it yourself with regexFind.
• OCR Extract Form Fields — extract by position instead of by pattern.