About the Remove Letters From Text
Run a scanned invoice or receipt through OCR and you get lines that mix a description with the figures you actually need. Deleting the letters leaves the numeric skeleton, which is often enough to eyeball the totals or paste into a sheet: a line reading Widget A x3 at 34.50 = 103.50 comes back as the quantities, the prices and the equals sign, in their original columns.
Removal works by Unicode letter category, so it is not limited to A to Z. Cyrillic, Greek, Arabic, Hebrew and Han characters all go, which makes the tool usable on multilingual paperwork where an ASCII-only filter would leave half the text sitting there. Digits, punctuation, symbols, currency signs and every kind of spacing stay exactly where they were, so aligned columns keep their alignment.
Combining marks are not letters, and that is the wrinkle. If a document stores é as an e plus a separate accent, the e is deleted and the accent is left behind, attaching itself to whatever character now precedes it. A pass through Remove Accents first avoids the stray marks. For a clean one-per-line list of the figures rather than a stripped layout, use Extract Numbers, and Remove Numbers for the reverse.
How to use
- Paste the text.
- Copy the result with letters removed.
- Use Extract Numbers for a clean list of the figures.
Common questions
- Are non-English letters removed?
- Yes, letters of any script are removed. Digits, punctuation and symbols remain.
- Is this the same as Extract Numbers?
- No. This keeps punctuation and layout; Extract Numbers lists just the numeric values.
- Are currency symbols kept?
- Yes, symbols are not letters, so € and $ remain.