Extracts all numeric, percentage, or currency runs from text using a global scan
(including numbers attached directly to text such as 28%YoY, $14.2M/qtr, up28%,
1.2M→3.4M, 5k users, 100GB storage), skipping alphanumeric names (B2B, COVID-19,
Web3, MP4, H2O) where a non-percentage/non-currency match is glued to a letter
(or a letter followed by -) on its left, and filters them with IsNumericLiteralToken
to exclude quarters, years, ISO dates, ratios, and ordinals/versions.
Extracts all numeric, percentage, or currency runs from
textusing a global scan (including numbers attached directly to text such as28%YoY,$14.2M/qtr,up28%,1.2M→3.4M,5k users,100GB storage), skipping alphanumeric names (B2B,COVID-19,Web3,MP4,H2O) where a non-percentage/non-currency match is glued to a letter (or a letter followed by-) on its left, and filters them withIsNumericLiteralTokento exclude quarters, years, ISO dates, ratios, and ordinals/versions.