Query Data Files
Open CSV, Parquet, JSON, Excel, SQLite and DuckDB files as tables in VS Code, and query them with SQL, a visual query builder or charts. Powered by DuckDB, so files of several gigabytes open in seconds.
Features
Open any data file as a table. Parquet, Excel, SQLite and DuckDB files open straight in the table. For CSV, TSV, JSON and JSON Lines, choose Open in Query Data Files from the editor title or the Explorer. Databases list their tables and workbooks their sheets.
Query with SQL. The open file is this, and other files are named by path: SELECT * FROM this JOIN 'customers.csv' USING (customer_id). Click a column name to sort, or right-click a value to filter to it; the SQL above the table updates, so you can keep editing it.
Column profiles under each header show a histogram, the number of distinct values and the share of NULLs.
Charts picked from your columns, grouped by DuckDB so they stay fast on millions of rows. Click a bar or point to see its rows.
Visual query builder. Build a query as a flowchart of steps (filter, join, group, pivot, window, formula and more) with a preview of every step. Flows are saved as .qflow.json files and can be copied as SQL at any time.
SQL files. Press Ctrl+Enter (Cmd+Enter on macOS) to run the statement under the cursor; the results open beside the editor. The Data Files view in the Explorer lists every data file with its tables and columns.
Compare with Git to see the rows added, removed and changed since an earlier commit.
Export results as CSV, Parquet, JSON or Excel, or copy them as a Markdown table.
Big files: 45 million rows (3 GB of CSV) open in a few seconds and scroll end to end.
Works offline and follows your theme. Queries can only read files in your workspace and files you open.
Requirements
- VS Code 1.90 or later, on Windows, macOS or Linux (x64 or Arm64).
Known Issues
- Excel support covers
.xlsx workbooks, not the older .xls format.
- Files with thousands of columns take several seconds to open.
- Files must be on this computer; URLs and cloud storage aren't supported.
- CSV lines longer than 8 MB can't be read.
- In an untrusted workspace, data files open but
.sql files and query flows don't run until you trust it.
Report issues on GitHub.
Release Notes
See CHANGELOG.md.
| |
