tpaw Docs

Visual EXPLAIN

Analyze PostgreSQL query performance with tpaw's interactive EXPLAIN ANALYZE plan tree.

Visual EXPLAIN parses the JSON output of a PostgreSQL EXPLAIN query and renders it as an interactive collapsible tree. Each node in the plan shows its type, cost estimates, row estimates, and — when using EXPLAIN ANALYZE — actual execution time and row counts. A color bar on each node's left edge indicates its relative cost within the plan.

Running EXPLAIN

In the query editor, click the Explain button in the toolbar, or prepend your query with EXPLAIN or EXPLAIN ANALYZE and run it normally. tpaw detects when a result contains a QUERY PLAN column and automatically switches to the visual plan view instead of the standard data grid.

For actual timing data, use:

EXPLAIN (ANALYZE, FORMAT JSON) SELECT ...

Without ANALYZE, the plan shows only estimated costs and rows.

Summary bar

At the top of the plan view, a summary bar shows:

FieldDescription
Total costThe planner's estimated total cost for the root node
PlanningTime spent planning the query (with ANALYZE only)
ExecutionTotal execution time in milliseconds (with ANALYZE only)

Plan tree

Each node in the tree is rendered as a card with:

Node header

  • Node Type — the operation name (e.g., Seq Scan, Hash Join, Index Scan)
  • Relation name — the table or index being accessed, with alias if different (e.g., orders (o))
  • A +/− toggle button for nodes with child nodes

Metrics badges

Each node card shows up to four badges:

BadgeDescription
costStartup cost .. Total cost (planner estimate)
rowsActual rows / Estimated rows (actual shown only with ANALYZE)
timeActual total time in ms (shown only with ANALYZE)
widthAverage output row width in bytes

Cost color bar

The left border of each card is color-coded by relative cost:

  • Green — cost is less than 33% of the plan's maximum node cost
  • Orange/peach — cost is between 33% and 66% of maximum
  • Red — cost is above 66% of maximum — the most expensive nodes in the plan

This makes it easy to identify the costliest operations at a glance.

Collapsing nodes

Click the − button on any node with children to collapse its subtree. Click + to expand it again. This is useful for large plans where you want to focus on a specific branch without scrolling through many nested child nodes.

Tips

  • Use EXPLAIN ANALYZE to get actual row counts alongside estimates — large discrepancies between estimated and actual rows often indicate stale table statistics (ANALYZE your tables).
  • Nodes at the bottom of the tree are executed first (leaves before parents). Start there when investigating slow plans.
  • High cost on a Seq Scan with many rows suggests a missing index on the filtered columns.

On this page