AgensGraph for Python
  • Docs home
  • AgensGraph 2.18
  • PyPI
/
  • Get started
    • Installation
      • Requirements
      • Optional extras
      • Check the installation
      • Install from source
    • Quickstart
      • Connect
      • Read a vertex
      • Read plain values
      • Pass parameters
      • Follow an edge
      • Return a path
      • Write
      • Combine SQL and Cypher
      • The same in awaiting code
      • Close the connection
      • Where to go next
  • Using the driver
    • Connecting
      • Opening a connection
      • Selecting a graph
      • Awaiting code
      • What the server can do
      • Closing
    • Running statements
      • execute_query()
      • Cypher and SQL together
      • execute() and cursors
      • Streaming a large result
      • What a write changed
      • Many statements at once
      • One statement per call
    • Parameters
      • Placeholders
      • How each value is sent
      • Where a parameter is allowed
      • Identifiers
    • Graph values
      • Vertices
      • Edges
      • Paths
      • Lists of elements
      • Identities
      • Equality, hashing, immutability
      • Two renderings
      • Numbers
      • Handing values to something that wants JSON
    • Transactions
      • The default: a transaction opens by itself
      • Transaction blocks
      • Autocommit
      • Isolation
      • Read-only transactions
      • Transaction ids
      • Two-phase commit
    • Other tools
      • The DB-API
      • SQLAlchemy
      • A psycopg connection the driver did not make
  • Working with data
    • Bulk loading
      • Before loading
      • Vertices
      • Edges
      • Writing what may already be there
      • Give the key a column of its own
      • From a DataFrame
      • Embeddings in bulk
      • Memory during a large read
    • DataFrames and Arrow
      • From a result
      • How values become columns
      • Larger than memory
      • In awaiting code
      • Without a library
      • Loading a frame
    • Embedding vectors
      • Setting up
      • Two places to keep an embedding
      • Writing a vector
      • Reading a vector
      • Indexing
      • Searching
      • Tuning a search
      • Reading the indexes back
      • Sparse vectors
    • LISTEN and NOTIFY
      • Listening
      • Announcing
      • Receiving: an iterator
      • Receiving: a handler
      • Announcing changes to the graph
  • Schema
    • Inspecting a graph
      • Which graph is read
      • Graphs
      • Labels
      • How many elements each label holds
      • Indexes
      • Constraints
      • Properties with a column of their own
      • Describing a graph
      • The label table
    • Declaring a schema
      • Before you start
      • Seeing the statements first
      • Labels
      • Indexes
      • Constraints
      • Removing what was not asked for
      • Statements without a connection
  • Running in production
    • Pooling
      • Creating and opening a pool
      • Borrowing a connection
      • A budget for the whole borrow
      • Hooks
      • After the server restarts
      • A statement interrupted mid-flight
      • Watching a pool
      • Without a pool: NullConnectionPool
      • Awaiting code
      • Sizing
      • Closing
    • Timeouts and cancellation
      • Deadlines
      • Cancelling a statement
      • When the network goes quiet
    • Retries
      • Conflicts are expected
      • Classifying a failure
      • A retry loop
      • Losing a connection while committing
    • Untrusted statements
      • Three conditions
      • 1. A role that cannot run programs on the host
      • 2. A read-only transaction
      • 3. One statement
      • A hint, not a boundary
      • Telling the model what went wrong
    • Errors
      • The driver’s own classes
      • What a failure prints as
      • Handing a failure to someone else
      • Whether to try again
    • Observability
      • Query logging
      • Tracing
      • Notices
      • Keep callbacks short
  • Reference
    • Connections
      • Connecting
      • Connection
      • Results
      • Server capabilities
    • Values
      • Graph values
      • Values you send
      • Numbers
      • JSON
      • Notifications
    • Pools, deadlines and retries
      • Pools
      • Deadlines
      • Retries
      • Commit outcome
    • Schema
      • Catalog rows
      • Declarations
      • Statements without a connection
    • Vectors
      • Values
      • Statements
    • Columnar export
      • Functions
      • Layout
    • Errors
      • Raised for the server
      • Raised by the driver
      • Functions
    • Utilities
      • Observability
      • Quoting
      • Adapters
      • Bulk loading and memory
      • DB-API
    • Compatibility
      • Python and its libraries
      • AgensGraph servers
      • pgvector
    • Changelog
      • 2.0.0
  1. AgensGraph for Python /
  2. Working with data
View Source Open in ChatGPT Open in Claude

Working with data#

Loading many elements at once, exporting results to DataFrames, embeddings, and notifications.

  • Bulk loading
  • DataFrames and Arrow
  • Embedding vectors
  • LISTEN and NOTIFY
Previous
Other tools
Next
Bulk loading

2026, SKAI Worldwide Co., Ltd

Made with Shibuya theme.