Unicode-aware counts, reading time and platform limits
Number Extractor
Extract every number from any text: Latin, Arabic-Indic and Persian digits, decimals with a dot or comma, percentages and scientific notation, with sum, mean, min, max and CSV export.
Workspace
Numbers
How the result is calculated
The text is scanned with a regular expression that matches a number (optional sign, Latin, ٠-٩ or ۰-۹ digits, . , ٫ ٬ separators, scientific exponent, %) and does not match digits glued to letters. Decimal separator: if both . and , appear the last one is decimal; a separator repeated is thousands; a single one followed by exactly 3 digits (1,234) is ambiguous, read as thousands by default and flagged; otherwise it is decimal.
Limits:
Local separator detection can be wrong for cases like
1.234 (thousands in German, decimal in English); set the option manually when needed. Phone numbers, dates and currencies are not extracted as entities (their digits are read as plain numbers).FAQ
How are ١٢٣٤ handled?
Arabic-Indic and Persian digits are converted to Latin in the value, and the original form stays in the raw column.
Why is 1,234 ambiguous?
It may mean one thousand two hundred thirty-four or 1.234; choose the reading.
