What Normalize Null Values changes
Choose the markers that represent missing data, then replace those markers with one selected value. You can set a table-wide target, give selected columns their own targets, review the preview, and export the normalized table as CSV.
How to normalize null values in a table
- Paste the table into Table data, or drop a CSV, TSV, or TXT file into the upload area and select browse files.
- Select the missing-value markers to recognize, then choose a target under Normalize missing values to.
- Leave First row is a header checked when the first row contains column names. Clear it when the file has no header; the tool then creates names such as
Column 1. - Review the detected columns. Enable Use a different representation for this column for any column that needs its own target.
- Select Export CSV in Preview to download the normalized table, or use the Export CSV in Change log to download the recorded changes.
Which missing-value markers can be normalized
Enable the controls that match your data: Empty string, Whitespace-only, NULL, N/A, NA, unknown, - (dash), or none. Add comma-separated values such as TBD or missing in Custom missing-value markers (comma-separated). Marker matching ignores letter case.
Under Normalize missing values to, choose Empty string, NULL, N/A, NA, - (dash), or Custom.... For example:
name,city,status
Acme,N/A,unknown
Beta,Chicago,-With Empty string selected, the result is:
name,city,status
Acme,,
Beta,Chicago,How per-column null value targets work
The global target applies to every column until you check Use a different representation for this column. That column then gets its own target selector, including Custom.... A status column can use NA while another column uses an empty string. Any value outside the selected markers stays unchanged.
CSV parsing, preview, and export rules
The parser detects comma, tab, semicolon, and pipe delimiters and reads quoted fields. For rows with different lengths, it uses the most common column count, pads short rows, and merges extra fields. With First row is a header unchecked, it creates names such as Column 1. Every normalized export includes a header row.
Preview displays the first 200 rows and highlights normalized cells. The visible Change log lists up to 500 changes; its count and CSV export include the full change list. The normalized export is comma-separated CSV, including when the input uses tabs, semicolons, or pipes. The page code contains no request that sends the table content elsewhere.
When to use a per-column target
Use one global target when the table follows one missing-value convention. If a status field must use NA while other fields use an empty string, enable its column override and check the highlighted result before exporting.
Choose the right dataset cleaning tool
Browse Dataset Clean Tools for related cleanup tasks. Choose Fill Blank Cells when blank cells need replacement values chosen by a rule. Use Data Cleaner for format-aware cleanup of CSV, TSV, Excel, JSON, JSONL, XML, or YAML input. Choose Number & Currency Normalizer for numeric or currency values.
Frequently asked questions
How do I normalize null values in CSV data?
Paste or upload the table, select the markers that count as missing, and choose a target under Normalize missing values to. Review the updated preview, then select Export CSV.
What counts as a null value in this tool?
The tool treats any enabled marker as missing, including NULL, N/A, NA, unknown, and - (dash). You can also enable empty or whitespace-only cells and add custom markers. Matching ignores letter case.
Can I use a different null value format per column?
Check Use a different representation for this column and choose that column's target. Other columns keep the global target.
Does the null value normalizer change every row?
It changes cells that match the selected markers. Other values stay unchanged. The preview highlights normalized cells, and the Change log records the old marker and new value.
Is the normalized null value export always CSV?
Yes. The normalized table downloads as comma-separated CSV, even when the input uses tabs, semicolons, or pipes.