SmartQueryTools

Aggregate Parquet Files Online

Group and aggregate Parquet files by any column directly in your browser. Calculate sum, average, min, max, and count for any numeric column — no upload required.

How to aggregate Parquet files

  1. Drop your file onto the upload area. The first column is chosen as the Group by column, and every numeric column is set to Sum.
  2. Pick the Group by column. The dropdown shows each column's type. One column is used, and each distinct value becomes one output row.
  3. For each other column, click one function: Sum, Avg, Min, Max or Count. Click the dash to leave a column out.
  4. Click Run Aggregation. The result shows the number of groups, sorted by the group value.
  5. Download the summary table in the same format as your file.

Your file is processed locally in your browser and is never uploaded. The free limit is 50 MB per file; larger files work if your device has the memory for them.

Worked example

A facilities team exports monthly electricity readings for three sites and wants one line per site for the quarterly report.

Input (Parquet)

Parquet file (binary, columnar) — shown as a table with its schema

sitemonthkwhcost
Depot North2026-011240.5310.13
Depot North2026-021102275.5
Head Office2026-01860.25215.06
Head Office2026-02912.75228.19
Depot South2026-01640160

Schema: site VARCHAR, month VARCHAR, kwh DOUBLE, cost DOUBLE

Settings

  • Group by: site
  • month: Count
  • kwh: Sum (the default for numeric columns)
  • cost: Avg

Result

sitemonth_countkwh_sumcost_avg
Depot North22342.5292.815
Depot South1640160
Head Office21773221.625

Five readings collapse to one row per site, sorted alphabetically. Each output column is named after its source and function, such as kwh_sum. Count on month counts the non-empty values, so Depot South shows one month of data. Avg is not rounded, so 292.815 keeps all its decimals. Run Round Numbers on the result if needed.

Working with Parquet files

Parquet files usually arrive with proper numeric types, so every INT, DECIMAL and DOUBLE column is set to Sum as soon as the file loads. Group by any column, including dates, timestamps and booleans. Grouping on a timestamp gives one row per distinct instant, which is rarely what you want. Run Date Truncate first to group by day or month.

Result types follow the function. Avg always returns a DOUBLE. Sum of a DECIMAL column stays DECIMAL, so money totals do not pick up floating-point noise. Sum of an integer column is widened to avoid overflow and is stored as a floating-point column in the Parquet output. Min and Max keep the source type. Count is always a whole number. The result is a small table with one row per group, written as a new Parquet file with the aliases shown in the preview.

Frequently Asked Questions

Can I group a Parquet file by month when it only has a timestamp column?

Not in one step. Run Date Truncate to add a month column, then aggregate with that column as Group by.

Will summing a DECIMAL column in Parquet lose precision?

No. Sum over a DECIMAL column keeps a DECIMAL result. Avg returns a DOUBLE, so round it if you need a fixed number of places.

Can I group by more than one column?

No. The tool groups by a single column. To group by two, combine them first with Combine Columns, then group by the new column. The SQL workspace supports any number of GROUP BY columns.

Can I apply more than one function to the same column?

No. Each column gets one function per run. To get both the sum and the average of a column, run the tool twice, or use the SQL workspace.

Does Count count rows or values?

Values. Count on a column counts the rows in each group where that column is not empty. Pick a column with no gaps, such as an ID, to get a row count.

Related Tools