Datui

Datui is a terminal UI for looking at tabular data: Parquet, CSV, JSON, Arrow and more, on disk or in S3, GCS and HTTP, from a few rows to a few billion.

datui data.parquet                    # open a file
datui --hive s3://bucket/warehouse/   # or a partitioned dataset in the cloud
datui                                 # or pick one from the home screen

Press ? for the keys. / queries, s sorts and filters, c charts, a analyzes, q quits. The documentation has the rest.

curl -fsSL https://raw.githubusercontent.com/derekwisong/datui/main/scripts/install/install.sh | sh

See the install guide for details and building from source.

• Opens the file where it is. Local paths, s3:// and gs:// URLs, and hive-partitioned directories; HTTP URLs are fetched first. Parquet is read lazily through Polars, one row group at a time, so a dataset larger than memory scrolls like a small one.
• Shows you what is around you. Run datui with no arguments for the home screen: recent datasets, the current directory, your data directories, and the buckets your credentials can reach. Each row shows rows, columns and size before you open it; the schema is one keystroke away.
• Answers questions. SQL, a short query language (select a, b where c > 10 by region), and fuzzy text search across every column. Sort, filter, freeze and hide columns from a sidebar.
• Summarizes. Describe, distribution fitting with Q-Q plots, and a correlation matrix, computed on the data as filtered.
• Draws. Line, scatter, bar, histogram, box, KDE and heatmap charts in the terminal, exportable as PNG or EPS.
• Reshapes and saves. Pivot and melt, export to CSV, Parquet, JSON, NDJSON, Arrow or Avro, and templates that replay a query, filters and sort on the next dataset with the same shape.
• Configurable. Light and dark palettes with every color a config key, defaults for every option, and arrow keys or h j k l.

import polars as pl
import datui

datui.view(pl.scan_parquet("data.parquet"))   # a LazyFrame stays lazy
datui.view("s3://bucket/events/", hive=True)  # paths and URLs work too

See the Python module.

datui --generate-config

Data directories for the home screen, S3 endpoints, CSV defaults, number formatting and every color live in one TOML file. See the configuration guide.

Bug reports and feature requests go to the issue tracker; security problems go to SECURITY.md instead. To build, test or contribute, start with the developer guide.

Datui is MIT licensed. See LICENSE.