Hive

Experimental

Creates:

AssetsLineage

Configure in the UI

This plugin can be configured directly in the Marmot UI with a step-by-step wizard.

View Guide

The Hive plugin discovers databases, tables, views and materialized views from Apache Hive through HiveServer2. It captures columns with their Hive types and comments, partition and bucket layout, storage formats, table statistics and view queries, and creates lineage from a database to its objects, from base tables to views and along foreign key constraints.

Every table and view is named database.table, so mrn://table/hive/sales.orders is the orders table in the sales database. Sample data previews are supported.

Connecting

The plugin connects to HiveServer2 over Thrift, on the binary transport (port 10000) or the HTTP transport (transport: http, usually port 10001 and path cliservice). auth selects the mechanism HiveServer2 is configured with: NONE (the default, PLAIN SASL without a password check), NOSASL, LDAP (username and password) or KERBEROS. Kerberos needs a plugin binary built with go build -tags kerberos, which links the GSSAPI library; the published binaries are built without it.

Row counts and sizes come from the metastore statistics, so they are as fresh as the last ANALYZE TABLE or stats-collecting write. Hive 4 creates non-ACID tables as external tables, and object_type reports what Hive reports.

Example Configuration

host: "hive.internal"
port: 10000
auth: "LDAP"
username: "marmot_reader"
password: "secret"
databases:
  - "sales"
  - "marketing"
include_ddl: false
tags:
  - "hive"
  - "warehouse"

Configuration

The following configuration options are available:

Property	Type	Required	Description
auth	string	false	Authentication mechanism (NONE, NOSASL, LDAP, KERBEROS)
databases	[]string	false	Databases to discover (all when empty)
exclude_databases	[]string	false	Databases to skip
external_links	[]ExternalLink	false	External links to show on all assets
filter	Filter	false	Filter discovered assets by name (regex)
host	string	true	HiveServer2 hostname or IP address
http_path	string	false	Endpoint path for the http transport
include_columns	bool	false	Include column information
include_ddl	bool	false	Store the SHOW CREATE TABLE output in metadata
include_partitions	bool	false	Count the partitions of partitioned tables
include_statistics	bool	false	Include row counts and sizes from table statistics
include_views	bool	false	Include views and materialized views
kerberos_service_name	string	false	Service part of the HiveServer2 Kerberos principal
password	string	false	Password, required for LDAP
port	int	false	HiveServer2 port
ssl	bool	false	Connect over TLS
ssl_skip_verify	bool	false	Skip TLS certificate verification
tags	TagsConfig	false	Tags to apply to discovered assets
transport	string	false	Thrift transport (binary or http)
username	string	false	Username (hive when empty)

Available Metadata

The following metadata fields are available:

Field	Type	Description
bucket_columns	[]string	Columns the table is bucketed by
column_name	string	Column name
comment	string	Database comment
comment	string	Table comment
compressed	bool	Whether the storage is marked compressed
created	string	Creation time (RFC 3339)
data_type	string	Hive data type as declared, complex types included
database	string	Database name
database	string	Database the object belongs to
ddl	string	SHOW CREATE TABLE output, when include_ddl is set
default_expression	string	Default value from a DEFAULT constraint
description	string	Column comment
hive_version	string	Hive version reported by the server
host	string	HiveServer2 hostname
input_format	string	Hadoop input format class
is_nullable	bool	False when a NOT NULL constraint covers the column
is_partition_column	bool	Whether the column is a partition column
is_primary_key	bool	Whether the column is part of the primary key constraint
last_access	string	Last access time (RFC 3339), when Hive tracks it
last_ddl	string	Time of the last DDL change (RFC 3339)
location	string	Warehouse location of the database
location	string	Storage location
num_buckets	int	Number of buckets, for bucketed tables
num_files	int64	Number of files, from table statistics
object_type	string	Object type (managed, external, view, materialized_view)
output_format	string	Hadoop output format class
owner	string	Database owner
owner	string	Object owner
owner_type	string	Owner type (USER, ROLE, GROUP)
parameters	map[string]string	Database properties (DBPROPERTIES)
parameters	map[string]string	Table properties not surfaced as their own field
partition_columns	[]string	Partition columns
partition_count	int	Number of partitions
partition_values	[]string	First 20 partition specs (dt=2026-01-01)
port	int	HiveServer2 port
serde	string	SerDe class
sort_columns	[]string	Columns each bucket is sorted by
table_count	int	Number of tables discovered in the database
table_name	string	Table or view name
transactional	bool	Whether the table is ACID (transactional)
view_count	int	Number of views and materialized views discovered in the database