Data Type Detector & Profiler

Analyze CSV/JSON columns to detect types (int, float, date, email, URL), null rates, and unique counts.

Tested tool guide Tested browser tools Checked August 16, 2026

What Data Type Detector & Profiler does and how it behaves

Paste CSV or JSON to see a column-by-column profile of the values present. The tool reports inferred labels including integer, float, date, email, and URL, together with missingness and distinct-value counts. This is exploratory inference, not a declared schema or proof that every value is valid. The common surprise is that inconsistent or ambiguous values can prevent a column from having the expected type, so inspect the source values before treating a detected label as authoritative.

How the result is produced

Column type inference

The pasted CSV or JSON is interpreted as fields that can be compared across records. CSV rows contribute values by column, while JSON records contribute values under recurring property names. The observed contents are assessed for recognizable numeric and structured-text forms. The resulting label summarizes what the column appears to contain; it does not convert or rewrite the input.

Completeness and cardinality

Alongside each type label, the profile reports a null rate and a unique count. The null rate summarizes values the tool recognizes as absent, while the unique count summarizes distinct values recognized within that column. Reading both measures together helps distinguish sparsely populated fields, repeated categories, and identifier-like columns without manually counting records.

Good uses

Inspect a newly received CSV before loading it into a database, especially when numeric columns may contain blanks or formatted text.
Compare exported JSON records with expected field shapes and locate columns whose dates, emails, or URLs do not profile consistently.
Screen candidate key and categorical columns by reviewing their unique counts together with missingness and inferred type.

Limits and checks

Type detection is inference from the pasted values, not validation against a data contract. A column labeled email or URL can still contain values unsuitable for the receiving application.
Null rate depends on which representations are treated as missing. Empty CSV fields, explicit JSON null values, whitespace, and text such as NA need not be equivalent, so check the original representation.
Unique count can be affected by representation. Case differences, surrounding spaces, numeric formatting, and date formatting may describe the same real-world value while still being counted separately.

Common questions

Does a detected date type prove that the dates were interpreted correctly?

No. A date label means the observed values were interpreted as date-like, but it cannot establish the intended locale, time zone, or business meaning. An input such as 03/04/2026 is ambiguous without a stated convention. Normalize dates to an unambiguous representation and inspect boundary cases before relying on the profile.

Does a unique count equal to the record count prove that a column is a primary key?

No. Matching counts, especially with a zero null rate, provide useful evidence of uniqueness in the pasted sample. They do not establish future non-nullability, stability, or business-level identity. A valid key rule must also come from the dataset's contract or producer, not solely from the profiler's observed counts.

References and verification

The behavioral notes were checked against the browser implementation. Standards and primary references below define the relevant format, formula, or platform behavior.