A CSV pivot table groups records into a smaller summary. Pivot Table Builder lets you choose the row field, column field, value field and aggregate. Decide what one output cell should mean before selecting those fields, because a count and a sum answer different questions.
Define one cell in plain language
Consider an illustrative CSV with columns team, month and amount. Its rows are A/Jan/10, B/Jan/20 and A/Feb/30. Choose team for rows, month for columns, amount for values and Sum for the aggregate. The output contains A with 10 in January and 30 in February, plus B with 20 in January.
| Team | January sum | February sum |
|---|---|---|
| A | 10 | 30 |
| B | 20 |
The B/February combination has no source record. A blank output cell indicates that missing combination; do not relabel it as an observed zero without checking what your dataset means.
Count would instead count rows in each group. Average, minimum and maximum use their own meanings. Write the selected aggregate in the table's caption or exported notes so another reader can interpret the cells.
Inspect the source before grouping it
The uploaded file must be UTF-8 CSV with a named header row. Duplicate or blank headings make column selection ambiguous and are refused. Use Dataset Profiler to inspect a source table's shape and candidate problems before reducing it to a summary.
Numeric aggregates require plain decimal values. Currency signs, grouped thousands such as 1,200 inside a cell, and unit labels are not silently converted into numbers. Correct the intended numeric spelling in a source copy rather than stripping characters without checking their meaning.
An optional filter matches the selected text without regard to case. Read the number of included source rows after filtering. Comparing a filtered pivot with an unfiltered total would create an apparent mismatch even when both calculations are correct.
Verify the export independently of the preview
The pivot preview can show only part of a large result. Download the CSV and inspect its full row and column set. The tool caps the pivot at two hundred rows and sixty columns; a very wide result needs narrower grouping or filtering.
For the illustrative Sum example, 10 plus 20 plus 30 gives a total of 60. Adding the visible populated pivot cells should return the same total. Other aggregate types do not necessarily add back to the source total.
Chart Builder from Data can visualize the resulting summary after the numbers are checked. Keep spreadsheet-formula protection in mind when inspecting exported labels; the export may prefix a formula-like label with an apostrophe so a spreadsheet displays it as text.