Analysis

Result management

dadbod-grip keeps result buffers alive as long as you want them. By default, running a new query replaces the existing grid. Pin a result to protect it from being replaced.

Pinning results

Press gL to pin the current result. Pinned results are never replaced by subsequent query executions. The buffer name gains a [pinned] suffix so you can tell them apart. Press gL again to unpin.

Pin multiple results to compare queries side by side. Run one query, pin it, run another, and both grids stay open in their own splits.

To cap the number of pinned results, set pinned_max in your setup:

require('dadbod-grip').setup({
  pinned_max = 3,
})

When the cap is reached, gL refuses to pin and notifies you to unpin something first.

Result switcher

Press gJ to open a picker over all live result buffers. Pinned results appear first. Selecting a result focuses its window, or opens it in the current window if it is not currently visible.

Use the result switcher to navigate a session where you have run several queries and want to return to an earlier result without hunting for the window.

Execute in new split

Press <S-CR> in the query pad to execute the current query and always open the result in a new horizontal split, regardless of whether an unpinned grid already exists. This is useful when you want to compare the output of two queries without pinning.

Table profile

Press gR or run :GripProfile to open a column distribution view. Each column shows a sparkline histogram, completeness percentage, cardinality estimate, and the top five most frequent values.

Use this to spot anomalies before writing queries: bimodal distributions, unexpectedly high NULL rates, or columns with a single value dominating the dataset.

Column statistics

Press gS to open a popup with detailed statistics for the column under the cursor:

• Total count and distinct count
• NULL count and null percentage
• Minimum and maximum values
• Top 10 most frequent values with counts

Aggregate on selection

Enter visual mode with V, select rows in a column, and press ga. A popup shows count, sum, average, minimum, and maximum for the selected cells.

Press ga without a selection to aggregate the entire current column across all pages.

Query Doctor

Run :GripExplain or press gQ to explain the current query. The output translates the raw EXPLAIN plan into plain-English health checks with cost bars and index suggestions.

Look for:

• Sequential scans on large tables (missing index)
• Nested loop joins on high-cardinality columns
• Sort steps without an index to satisfy the ORDER BY

Row diff

Select exactly 2 rows in visual mode and press gd. Cells that differ between the two rows highlight in amber. Matching cells are unchanged. Use this to spot what changed between two versions of the same record, or to compare rows that should be identical.

Three comparison operations cover different needs:

Data diff

Press gD or run :GripDiff to compare two tables by primary key. The grid shows rows that differ between the two tables, color-coded by the type of change.

The diff works across databases when running through a DuckDB federation session.

Filter builder

Press gF to open an interactive filter builder. Choose a column, select an operator, and enter a value. Multiple filters stack with AND.

Press f for a faster shortcut: it filters the current column to the exact value under the cursor in one keystroke. Press F to clear all active filters.

Saved filter presets

Save the current filter combination as a named preset with gP. A prompt asks for a preset name. The preset captures all active filters, including column, operator, and value for each one.

Load a saved preset with gp. The picker shows all presets for the current table with a preview of the filter conditions. Selecting a preset replaces the current filters entirely.

Presets persist in .grip/filter_presets.json relative to the current working directory. They are scoped to the table name, so presets saved on the orders table only appear when viewing orders.

Data diff across databases

gD compares two tables by primary key. In a DuckDB federation session, you can diff a Postgres table against an archive SQLite table:

1. Open a federation session with both databases attached
2. Open the Postgres table in a grid
3. Press gD and select the SQLite table as the comparison target

Rows that differ appear color-coded by change type. Useful for migration validation and schema drift detection.

Export

Press gE to export the current result set to the clipboard in one of six formats: CSV, TSV, JSON, SQL INSERT, Markdown pipe table, or Grip Table (box-drawing borders).

Press gX or run :GripExport to write to a file. Choose CSV, JSON, or SQL INSERT. The file goes to the current working directory by default.

Sort stacking

Press s to toggle sort on the current column (ASC then DESC then off). Press S to add a secondary sort without replacing the first. Columns show their direction and stack position:

 id  ▲1  name  ▼2  created_at

This grid is sorted by id ascending first, then by name descending within ties. Stack up to as many columns as needed. Press X to clear all sorts, filters, and pagination in one keystroke.