Build a CSV pivot table

Choose group columns and the column supplying categories. Count records or sum a selected decimal column.

  • Runs in your browser
  • No upload
  • Check before download

Files and cell values stay in this tab. Nothing is uploaded or saved. Closing or clearing the tab loses the inputs.

Choose group columns and the column supplying categories. Count records or sum a selected decimal column. Choose the empty-cell rule; keep means zero for an empty sum value. Original sources are unchanged, and the output shows one column per category.

UTF-8 only; BOM accepted. Cells and keys remain exact strings: 001 differs from 1. An empty cell is empty text, not inferred null; literal NULL stays text. Headers must be unique and nonempty. Quoted newlines are preserved. Formula-like cells remain unchanged; spreadsheets may execute them on opening.

First record is a header

Transforms: 8 files, 16 MiB per file and 32 MiB in total, 200,000 records, 128 columns, and 1,000,000 input or output cells. Split: 64 MiB, 1,000,000 records and 500 parts. A field can have up to 262,144 UTF-16 units and a record up to 1,048,576 characters. Output up to 128 MiB; the worker stops after 60 seconds. The ZIP bytes stay in memory, so this is not an unlimited disk stream.

Inspect the selected sources before choosing columns.

Output

How to use it

  1. Choose files and delimiters

    UTF-8 only; BOM accepted. Cells and keys remain exact strings: 001 differs from 1. An empty cell is empty text, not inferred null; literal NULL stays text. Headers must be unique and nonempty. Quoted newlines are preserved. Formula-like cells remain unchanged; spreadsheets may execute them on opening.

  2. Inspect and configure columns

    Choose group columns and the column supplying categories. Count records or sum a selected decimal column. Choose the empty-cell rule; keep means zero for an empty sum value. Original sources are unchanged, and the output shows one column per category.

  3. Create and verify the native download

    Sum accepts signed decimal text with a dot, up to 18 integer and 12 fractional digits. No exponent, grouping separator or automatic coercion. Absent combinations export zero; category order follows first occurrence.

Before exporting

CSV files

Files and cell values stay in this tab. Nothing is uploaded or saved. Closing or clearing the tab loses the inputs.

Empty-cell policy

UTF-8 only; BOM accepted. Cells and keys remain exact strings: 001 differs from 1. An empty cell is empty text, not inferred null; literal NULL stays text. Headers must be unique and nonempty. Quoted newlines are preserved. Formula-like cells remain unchanged; spreadsheets may execute them on opening.

Output

Transforms: 8 files, 16 MiB per file and 32 MiB in total, 200,000 records, 128 columns, and 1,000,000 input or output cells. Split: 64 MiB, 1,000,000 records and 500 parts. A field can have up to 262,144 UTF-16 units and a record up to 1,048,576 characters. Output up to 128 MiB; the worker stops after 60 seconds. The ZIP bytes stay in memory, so this is not an unlimited disk stream.