Skip to content

TurboLynx CLI — Overview

The turbolynx binary is the all-in-one command-line interface for TurboLynx. It exposes two subcommands:

Subcommand Description
turbolynx shell Interactive Cypher query shell (default)
turbolynx import Bulk-load graph data from CSV files

Build instructions live in the installation guide. Native binaries are platform-specific: Linux fast-path builds produce ./build/tools/turbolynx, while the portable Linux/macOS build produces ./build-portable/tools/turbolynx.

The native shell can execute both read and write Cypher queries. CREATE, SET, DELETE, DETACH DELETE, and basic node MERGE go through the same mutation path used by the C API and Python API.

Starting the Shell

turbolynx shell --workspace /path/to/db

If no subcommand is given, shell is assumed:

turbolynx --workspace /path/to/db

The shell opens an interactive REPL:

TurboLynx shell — type '.help' for commands, ':exit' to quit
TurboLynx >>

Shell Options

Flag Short Description
--workspace <path> -w Path to the database directory (required)
--query <string> -q Execute a single query and exit (non-interactive)
--mode <name> -m Set output format at startup (see Output Formats)
--iterations <n> -i Repeat each query N times (benchmarking)
--warmup Discard the first iteration from timing statistics
--profile Enable query profiling
--explain Print the selected physical plan (no execution)
--compile-only -c Parse and plan without executing
--standalone -S Accepted for compatibility; currently has no effect in the native CLI
--log-level <level> -L Logging verbosity (trace, debug, info, warn, error)

Running a file of queries. There is no --query-file flag yet. To execute a script of Cypher statements, start the shell and use the .read <file> dot command, which runs each statement in the file in order.

Toggling profile / explain at runtime. --profile can be turned on and off inside an interactive session with .profile on / .profile off. Explain mode currently has no dot-command equivalent — to switch in and out of explain, restart the shell with or without --explain.

Join Order Options

Flag Description
--join-order-optimizer <type> query · greedy · exhaustive · exhaustive2 · gem
--index-join-only Force index nested-loop join
--hash-join-only Force hash join
--merge-join-only Force merge join
--disable-merge-join Disable merge join
--disable-index-join Disable index join
--debug-orca Print ORCA optimizer debug output

Non-Interactive Mode

Run a single query and write results to stdout:

turbolynx shell --workspace /path/to/db \
                --query "MATCH (n:Person) RETURN n.firstName LIMIT 5;"

Combine with --mode to control output format:

turbolynx shell --workspace /path/to/db \
                --mode csv \
                --query "MATCH (n:Person) RETURN n.firstName, n.lastName;"

Entering Queries

  • Cypher queries must end with a semicolon (;).
  • Multi-line input is supported — the shell continues reading until it finds ;.
  • Dot commands (.mode, .tables, etc.) do not require a semicolon.
TurboLynx >> MATCH (n:Person)
TurboLynx -> RETURN n.firstName
TurboLynx -> LIMIT 10;

Keyboard Shortcuts

Key Action
Tab Autocomplete dot commands, Cypher keywords, labels, and edge types
Ctrl+D Exit the shell (EOF)
Ctrl+L Clear the screen
↑ / ↓ Navigate command history

Exiting

TurboLynx >> :exit
TurboLynx >> .exit
TurboLynx >> .quit

Or press Ctrl+D.

Import Subcommand

See Data Import for full import documentation.

turbolynx import \
    --workspace /path/to/db \
    --nodes Person dynamic/Person.csv \
    --relationships KNOWS dynamic/Person_knows_Person.csv
Flag Description
--workspace <path> Output database directory
--nodes <label> <file> Vertex CSV file (repeatable)
--relationships <type> <file> Edge CSV file (repeatable)
--incremental <true\|false> Append edges to existing database
--skip-histogram Skip histogram generation after load
--log-level <level> Logging verbosity