CSV data quality
Blank vs Zero vs Null in CSV Files: How to Analyze Missing Values Safely
A blank CSV field, a numeric zero, and a null value are not automatically the same. Zero is a known numeric value. A blank field contains no visible characters, but its meaning depends on the source: it may mean missing, not applicable, not yet entered, deliberately withheld, or an empty string. “Null” is a database or application concept for an absent value; CSV itself has no universal null type, so a source may export null as an empty field or as text such as NULL, N/A, or -. Before calculating totals, averages, rates, or charts, document the source convention, preserve the original file, normalize only confirmed missing-value markers, and report how missing records were handled.
Short answer
A blank CSV field, a numeric zero, and a null value are not automatically the same. Zero is a known numeric value. A blank field contains no visible characters, but its meaning depends on the source: it may mean missing, not applicable, not yet entered, deliberately withheld, or an empty string. “Null” is a database or application concept for an absent value; CSV itself has no universal null type, so a source may export null as an empty field or as text such as NULL, N/A, or -. Before calculating totals, averages, rates, or charts, document the source convention, preserve the original file, normalize only confirmed missing-value markers, and report how missing records were handled.
Upload
Bring the CSV you already use.
Understand
Start with the brief and evidence.
Act
Ask, edit, filter, and export.
A practical starting point
A blank CSV field, a numeric zero, and a null value are not automatically the same. Zero is a known numeric value. A blank field contains no visible characters, but its meaning depends on the source: it may mean missing, not applicable, not yet entered, deliberately withheld, or an empty string. “Null” is a database or application concept for an absent value; CSV itself has no universal null type, so a source may export null as an empty field or as text such as NULL, N/A, or -. Before calculating totals, averages, rates, or charts, document the source convention, preserve the original file, normalize only confirmed missing-value markers, and report how missing records were handled.
Concrete examples
Example 1: Zero sales is not a missing sales record
A daily outlet file contains sales values of 1250.00, 0.00, and an empty field. The zero means the outlet traded but recorded no sales. The empty field means the report did not provide a value. The dashboard keeps zero in the total and average denominator, while the missing record is counted and investigated separately.
Example 2: Blank discount has a documented default
An order system defines a blank discountamount as “no discount applied.” The analyst may normalize that field to zero because the source specification confirms the rule. The same conversion is not applied to blank ordertotal values, which remain missing and are excluded from revenue calculations pending correction.
Example 3: N/A means not applicable, not unknown
An inventory export uses N/A for expirydate on non-perishable products and leaves the field blank when an expiry date is missing for a perishable product. Combining both states as null would hide a data-quality problem. The working dataset keeps an applicability flag and reports the missing perishable dates separately.
Example 4: A parse failure creates a deceptive blank
A source value contains 1.234,56, but an import configured for another numeric convention fails to parse it and produces an empty cell. The analyst checks the raw CSV and import log before calling it missing. The value is corrected only after confirming the source locale and decimal convention.
Examples are illustrative and are not customer results.
Common questions
What is the safest way to work with this CSV problem?
A blank CSV field, a numeric zero, and a null value are not automatically the same. Zero is a known numeric value. A blank field contains no visible characters, but its meaning depends on the source: it may mean missing, not applicable, not yet entered, deliberately withheld, or an empty string. “Null” is a database or application concept for an absent value; CSV itself has no universal null type, so a source may export null as an empty field or as text such as NULL, N/A, or -. Before calculating totals, averages, rates, or charts, document the source convention, preserve the original file, normalize only confirmed missing-value markers, and report how missing records were handled.
Keep exploring
See what your own CSV is saying.
Try the browser workspace with a real file and keep the public guide open when you need help.
Start free