Arrow Viewer
Open Apache Arrow (.arrow), Feather (.feather) and IPC (.ipc) files in your browser. Browse the schema and page through the records in a fast table, with nothing uploaded.
Free, no sign-up, and fully client-side. Your file never leaves your browser. Opens .arrow, .feather, .ipc.
Open the viewer and drop a file to get started.
What is an Arrow file?
Apache Arrow is a columnar in-memory format for analytics, and its IPC and Feather files are that same layout written to disk. Because the on-disk format mirrors memory, Arrow files load with essentially zero parsing, which makes them popular for fast interchange between pandas, Polars, DuckDB and R.
ViewKit reads the Arrow IPC stream in your browser, exposes the schema with each field’s type, and renders the rows in a virtualized table. Nothing is uploaded.
What you can do with the Arrow viewer
Schema with types
See every field with its Arrow type, including lists, structs, dictionaries and timestamps.
Feather & IPC
Opens .arrow streams, .feather files and .ipc files, the same Arrow IPC format under different names.
Virtualized table
Records are windowed into the table, so wide, many-row files scroll smoothly.
Nested values
List and struct values render as JSON, and columns resize to fit their content.
How to open a Arrow file
- 1
Drop your .arrow, .feather or .ipc file onto the box above (or click to browse). It is read in your browser, nothing uploads.
- 2
Browse the schema to see each field and its type.
- 3
Page through the record table; resize columns to fit as needed.
Doing the same thing in code
The same file in Python, using pyarrow:
import pyarrow.feather as feather
table = feather.read_table("data.arrow")
print(table.schema) # fields and Arrow types
print(table.num_rows, table.num_columns)
print(table.slice(0, 10).to_pandas())The function to call depends on whether the file was written as a stream or as a file, and .arrow, .feather and .ipc are the same format under different names. This page opens all three the same way.
Arrow viewer: frequently asked questions
Does it open Feather files?
Yes. Feather is the Arrow IPC file format, so .feather, .arrow and .ipc all open the same way.
What is the difference between .arrow, .feather and .ipc?
They are the same format. All three are Arrow IPC, the in-memory columnar layout written to disk, and Feather v2 is Arrow IPC, which is why the extensions are interchangeable. The only real distinction is stream framing versus file framing, and both are handled. Feather v1 is an older, unrelated layout and is not supported.
Why does an Arrow file open faster than a Parquet file?
Arrow on disk is already the in-memory layout, so it can be read without decoding. Parquet is compressed and encoded for storage, so reading it requires decompression and decoding first. The trade-off is size: the same data is usually several times larger as Arrow than as Parquet, which is why Parquet is used for storage and Arrow for interchange.
Are dictionary-encoded columns supported?
Yes. Dictionary columns, which Arrow uses for repeated strings and is what a pandas categorical becomes, are decoded to their values so the table shows the actual strings rather than the integer codes.
My Arrow file will not open, why?
Files written with body compression (LZ4 or ZSTD) are not yet supported in the browser. Re-export the file uncompressed and it will open.
Is my data uploaded?
No. The file is parsed entirely in your browser; its contents never leave your machine.