Choose a data file
Open a local flat Parquet, Avro OCF, Arrow IPC File or Feather V2 table up to 20 MB.
Inspect bounded analytical file rows and schema in your browser.
Parquet Viewer
Open a bounded Parquet file locally to inspect rows, column schema, nullability and row groups without uploading data.
Flat scalar columns only; exact DECIMAL, nested and binary types are not claimed.
Data File ViewerYOUR WORKSPACE
Your task
View data
Choose a file to inspect it here.
Also with this file: Download CSV · Download Excel · Download JSON
Processed privately in your browserNo uploadNo signup
A simple workflow
Three steps. One less thing to do.
Review the data
Inspect physical and logical column types, row count, nullability and filtered, paginated rows.
Export a bounded table
Download the accepted rows as CSV, JSON or an Excel workbook with a schema sheet.
Made for the everyday
A useful part of your workflow.
Dataset triage
Check a Parquet schema and row sample before moving a data file into a pipeline.
Spreadsheet handoff
Export a bounded flat table with an accompanying schema sheet.
Null and type review
Identify nullable columns and compare source types with displayed values.
Good to know
A few useful answers.
The details that matter before you use Data File Viewer.
Which Parquet files work?
Parquet retains its existing bounded flat scalar scope. Avro OCF accepts uncompressed flat record blocks with null, boolean, int, long, float, double and string. Arrow IPC File and Feather V2 accept null, boolean, integer, float and UTF-8 columns. Unsupported structures fail before export.
Are large integers rounded?
No. BigInt values are exported as decimal strings. Unsafe JavaScript integer values are rejected rather than silently rounded.
Can I run SQL?
No. This viewer provides local row filtering and schema inspection. SQL is withheld until a browser query engine passes a separate size and fidelity review.
What do the downloads contain?
CSV and JSON contain the accepted rows. Excel adds a Schema sheet. CSV protects formula-like cells; Excel uses text cells.
Does the file leave my device?
No. The reader uses local file slices; row values, schema names and filenames are excluded from analytics.
What are the limits?
One file up to 20 MB, 20,000 rows, 128 columns and 200,000 cells, with up to 5 million decoded characters. Avro accepts up to 128 blocks; existing Parquet keeps its 128 row-group limit. Larger files are refused.
Are ORC, Feather V1 and Arrow streams supported?
No. ORC, Feather V1 and Arrow streaming IPC are withheld; this version supports Avro OCF, Arrow IPC File and Feather V2 within their stated scalar limits.