Skip to content
Queria Docs
English
日本語English
Esc
↑↓navigate↵open⌘Jpreview
On this page

View table details

How to use the tabs and features of the table detail page

Selecting a table opens its detail page, where you can inspect the table’s structure, column information, statistics, and data lineage.

Overview tab

Shows the table’s metadata.

  • Table description
  • Database name, schema name, table name
  • Kind (table / view)
  • License, data source URL, and source code (a link to the transformation logic)
  • Table definition (the CREATE TABLE / CREATE VIEW DDL)

Click the database name to copy it to the clipboard.

Columns tab

Lists the table’s column information.

  • Column name (plus its title, when one is set)
  • Data type
  • Whether a NOT NULL constraint applies
  • Column description

Statistics tab

Shows a statistical summary for the whole table and for each column.

Summary

Three metrics appear at the top of the page.

  • Rows: total row count of the table
  • Columns: total column count
  • Missing rate: the share of NULL values (missing cells out of all cells)

For large tables, a note indicates that the statistics are based on sampling.

Per-column statistics

Different statistics are shown depending on each column’s data type. The NULL rate is shown for every column.

Numeric types (INTEGER, FLOAT, DECIMAL, etc.):

  • Minimum, maximum, mean, median
  • Standard deviation, first quartile (Q1), third quartile (Q3)
  • Histogram: the value distribution as a bar chart
  • Box plot: visualizes the min–max range, the interquartile range (Q1–Q3), and the median

Categorical types (VARCHAR, TEXT, etc.):

  • Unique count and cardinality level (unique / high / medium / low)
  • Most frequent value and its share
  • Top 10 distribution: the share of the top 10 values as a horizontal bar chart, with the rest grouped as “other”

Date types (DATE, TIMESTAMP, etc.):

  • Earliest and latest date
  • Span in days
  • Number of unique dates

Boolean type (BOOLEAN):

  • Count and share of TRUE and FALSE

Guide tab

Appears only when recipes covering this table are available. The SQL shown can be edited and run right on the page. The same content is also collected in the cookbook.

Lineage tab

Shows the table’s lineage graph — a visual view of which tables it derives from and which tables use it.

  • The current table is highlighted with a thick border
  • Each node shows the table name and kind (table, view, ephemeral)
  • The layout direction can be toggled between horizontal and vertical
  • Full-screen mode is available (press Escape to exit)

Clicking a node opens a detail panel with that table’s description, SQL definition, column list, and lists of its upstream and downstream tables.

The lineage tab appears only for remote datasets (datasets with an active catalog).

History tab

Shows the snapshots in which this table was updated, newest first, like a commit history. Each entry shows a version number (for example v399), the update time, and the kind of change (added / updated / removed).

Select “View this version” to preview the table as it was at that version (time travel). The preview is set up with an AT (VERSION => n) clause in the SQL, and you can use the same syntax to run any query against a past version. The selected version is reflected in the URL, so sharing the link reproduces the same view.

Only snapshots retained in the catalog (about 7 days) are shown. Data from older versions is no longer available.

The history tab appears only for remote datasets.

Use the data

The “Use this data” button at the top right of the page lets you export the current table or open it in external tools.

Export:

  • Download as CSV, JSON, Parquet, or Excel
  • Choosing “Google Sheets” creates a spreadsheet directly in your Google Drive
  • Conversion runs in DuckDB inside your browser; nothing goes through a server
  • Choosing JSON for a table with a geometry column outputs GeoJSON
  • See Download the data for details

Open in a tool:

  • Queria CLI: copy a uvx queria command that reads this table. It bundles a compatible version of DuckDB, so it runs without installation, and referenced datasets are attached automatically. See the CLI guide for usage
  • DuckDB CLI: copy a duckdb -c command that attaches the DuckLake catalog and queries this table. Paste it into your terminal to run as-is (requires DuckDB 1.5.2 or later). See Connect from DuckDB CLI for details
  • Claude Code / Codex: copy a command that asks the agent to analyze this table. Paste it into your terminal and the agent receives a prompt that connects through the Queria CLI (uvx queria) and starts analyzing. With the Queria skill installed, the agent can explore across the whole catalog, including cross-dataset search. See Analyze with AI agents for the analysis workflow
  • Tableau Desktop: copy the initial SQL that attaches the DuckLake catalog. See Connect from Tableau Desktop for installing the connector and connecting

Preview tab

A SQL editor and its results. It starts with SELECT * FROM ... LIMIT 100, which runs automatically when you open the tab.

For tables with a geometry column (such as boundary data or facility locations), the preview tab shows a “Map | Table” switch and opens with a map preview by default.

  • Points, lines, and polygons are all drawn on the map, which zooms automatically to fit the data
  • The map shows a random sample of 500 geometries; the total count appears below the map
  • The switch state is reflected in the URL, so you can share a link that opens in map view

Edit the SQL editor to run your own queries.

  • To run: click the run button, or press Cmd+Enter (macOS) / Ctrl+Enter (Windows/Linux)
  • Results are shown as a table and can be exported

For aggregations and queries across multiple datasets, the sample SQL in the DuckDB CLI guide applies as-is — the same SQL runs in the browser preview too.

Next steps

Was this page helpful?