How to Profile a CSV File Before You Import It

On this page

A CSV can look normal and still contain surprises. A number column may include a few words. An email column may have empty cells. An ID column may contain more repeated values than you expected.

The CSV profiler gives you a quick picture of each column before you import, clean, or share the file.

What the profiler tells you

For each column, the report can show things such as:

how many cells are empty,
what kind of values the column mostly contains,
whether different kinds of values are mixed together,
the smallest, largest, and average number when the whole column is numeric,
the shortest and longest text,
how many different values appear,
which values appear most often when they can be counted exactly.

This is useful when you want to understand the data before another system has a chance to reject it.

When profiling is useful

Profile a CSV when you are about to import customer records, orders, survey results, product data, or any other table you did not create by hand.

It is especially helpful when a file came from an export, another team, an older system, or a spreadsheet with many manual edits.

If the CSV is already visibly broken, with rows shifted into the wrong columns or strange characters in the text, fix those problems first. Use the CSV validator for structural problems or the encoding guide for broken characters.

How to profile a CSV

01
Add your CSV
Choose one CSV. Your file stays on your device; the tool page shows the current file-size limit.
02
Choose how many unique values to count
The default works well for most files. Raise it only when an exact unique-value count matters for a column with many different values.
03
Run the profile
The tool reads the file and builds a summary for every column.
04
Look for surprising columns
Pay attention to empty cells, mixed values, unexpected ranges, and repeated IDs or codes.
05
Save the report if you need it
Download the profile report when you want to keep a record or compare the file with a later export.

How to read common results

What you see	What it may mean	What to do next
Many empty cells	Important values may be missing.	Check whether those blanks are allowed before importing.
A number column is marked as mixed	Some cells contain text such as N/A or unknown.	Find those cells and decide whether to correct, remove, or keep them.
An ID column has fewer different values than rows	Some IDs are repeated.	Check whether duplicates are expected.
A date column looks like text	The dates may use a format the profiler does not recognize as a clear date.	Check the date format required by the system you will import into.
A number range looks wrong	There may be a typo, wrong unit, or unexpected value.	Inspect the smallest and largest values before importing.

What happens to your file?

The CSV is processed on your device. It is not uploaded to our servers.

The profiler does not replace or edit the original file. It creates a separate report about what it found.

This makes it useful for private customer lists, financial exports, internal reports, and other files you would rather not upload.

Practical limits

The profiler accepts one CSV up to 200 MB. The cap follows the same registry method: the displayed limit comes from the shipped tool registry and is backed by first-party browser measurement. The first-party measurement evidence is recorded in docs/tool-audit/batch-a.md.

Unique-value counts also have a limit that you can choose. If a column has more different values than that setting, the report tells you that the real count is at least that high. It does not show a partial count as if it were exact.

Large files can take longer on older or slower devices.

Simple example

Imagine an amount column with 10,000 prices and one cell containing N/A.

The column may look numeric when you scan it in a spreadsheet, but the profiler shows that the values are mixed. That tells you to inspect the exception before sending the file to a system that expects a number in every row.

After: you know which column needs attention before the import starts.

What to do after profiling

If the report shows mixed values or unexpected blanks, correct the source data when possible and profile the new export again.

If rows have the wrong number of columns, headers are duplicated, quotes are broken, or the file has other CSV structure problems, use the CSV validator. If you need a cleaned copy, use the CSV cleaner.

Profile a CSV

See the transformation