Number & Currency Normalizer

How to use Number & Currency Normalizer

1. Drop a CSV, TSV, JSON, or text file here, or paste the table into Table data.
2. Select at least one column under Columns to normalize. Select all and Select none change the selection for every column.
3. Choose an Input separator / locale rule and an Output decimal precision.
4. Decide whether to check Retain original column(s) and Write normalized amount and currency-code columns.
5. Review Normalized preview, then choose Download normalized table (CSV), Download conversion report (CSV), or Download conversion report (JSON).

How separator rules read a number

Input separator / locale rule includes Auto-detect per value (flags ambiguous cases), US / UK, EU, Space-grouped, Swiss, and Custom... choices. Auto-detect marks a value ambiguous when one separator is followed by exactly three digits. The value 1,234 stays flagged until you choose an explicit rule.

price
"$1,234.50"
(450.00 EUR)

normalized_amount | currency_code | status
1234.50 | USD | ok
-450.00 | EUR | ok

The parser removes a recognized currency symbol or ISO code from the numeric text. When Write normalized amount and currency-code columns is checked, it writes the code beside the normalized amount. Parentheses and leading or trailing minus signs indicate a negative value. The amount stays in its source currency.

What Output decimal precision changes

Output decimal precision can Preserve source precision or format recognized amounts with 0, 2, 3, 4, or 6 decimal places. For example, 12.5 becomes 12.50 at precision 2. Ambiguous, invalid, and empty values receive no normalized amount.

When Retain original column(s) is checked, selected source columns remain beside the result. Clear it to remove them from the preview and CSV export. Write normalized amount and currency-code columns adds amount and code fields; status and notes remain for selected columns.

What happens to invalid or empty values

Delimited input needs a header row. The first line is used to choose comma, tab, semicolon, or pipe parsing. Quoted fields can contain separators or line breaks. JSON must be an array of records or an object containing a rows or data array. Missing cells become empty values.

Unrecognized characters in the numeric part produce invalid status, and incorrect thousands grouping is rejected under the selected rule. Normalized preview shows the first 500 parsed rows when the input is larger; downloads include every parsed row. FileReader reads selected files in the browser, and the normalization code does not send table contents to a server.

Choose a rule when formats differ

For 1,234.50, choose the US / UK rule. For 1.234,50, choose the EU rule. Space-grouped and Swiss rules cover their named separators. Auto-detect can read unambiguous values, but 1,234 needs an explicit choice.

Values such as $1,234.50 or (450.00 EUR) can produce a normalized amount and currency code. Retain original column(s) keeps the source text in the row.

For wider cleanup, start in Dataset Clean Tools. Use Data Cleaner for format-aware cleanup, Normalize Null Values for missing-value markers, or Value Clusterer & Standardizer for text labels that need review and approval.

Frequently asked questions