Hive Server 2 entities created in Atlas

Each Hive Server 2 entity in Atlas includes detailed metadata collected from Hive.

The following diagrams show a summary of the entities created in Atlas for Hive operations and assets. The supertypes that contribute attributes to the entity types are shaded.

Figure 1. Atlas Entity Types for Hive Server 2 Data Sets


Figure 2. Atlas Entity Types for Hive Server 2 Processes


The metadata collected for each entity type is as follows:

Hive Process

Identifier Example content
typeName hive_process
guid System generated ID. This value is used to identify the entity in the Atlas Dashboard URL.
qualifiedName

<database>.<target table>@<clustername>:<generated ID>

The generated ID is distinct from the GUID.

name Text of the query.
inputs List of the input tables or views, including each entity’s type name and the qualified name.
outputs List of the output objects, including each entity’s type name and the qualified name.
recentQueries Last query executed (duplicated in process_execution).
operationType One of the operations that triggers metadata collection.
queryPlan Reserved for future use.

Hive Process Execution

Identifier Example Content
typeName hive_process_execution
guid System generated ID. This value is used to identify the entity in the Atlas Dashboard URL.
qualifiedName <database>.<target table>@<clustername>:<ID from process qualified name>:<ID from the process execution name>:<generated ID for this process execution>
name Text of the query with a system-generated ID added to the end.
queryText Text of the query.
queryPlan Reserved for future use.
queryId impala_<date as yyyymmddhhmmss>_<generated id>
startTime Query start time.
endTime Query end time.
userName The user who ran the query.
Relationship: Process One process to one or more process executions. hive_process_process_execution

Hive Database

Identifier Example Content
typeName hive_db
guid System generated ID. This value is used to identify the entity in the Atlas Dashboard URL.
qualifiedName <database>@<clustername>
name Database name as reported from Hive.
clusterName Cluster name.
location The file system path where the backing files for the database are stored. This could be an HDFS path, an AWS S3 object, or an Azure data storage location.
owner The user who initially created the database.
ownerType The principal type of the database owner. Could be USER, ROLE, or GROUP.
parameters Additional key-value pair metadata that comes from Hive such as table size, number of rows, and number of storage files.
Relationship: Table One database to many tables. hive_table_db
Relationship: Database DDL One database to many database DDL entities. hive_db_ddl_queries

Hive Table

Identifier Example Content
typeName hive_table
guid System generated ID. This value is used to identify the entity in the Atlas Dashboard URL.
qualifiedName <database>.<tablename>@<clustername>
name Table name.
columns List of the columns defined in the table. The Atlas Dashboard shows these as links to the column entity details.
owner The user who created the table.
parameters

Table details from Hive Server 2 such as:

  • totalSize

  • External

  • numFiles

  • transient_lastDdlTime

  • bucketing_version

retention Provided by HS2. Integer value
sd

The location of the table data, the storage description.

<database>.<table>@<clustername>_storage

tableType How the table was created: one of EXTERNAL_TABLE, VIRTUAL_VIEW, or MANAGED_TABLE.
Relationship: Database One database to many tables. hive_table_db
Relationship: Columns One table to one or more columns. hive_table_columns
Relationship: Partition Key Column One table to one or more columns that are partition keys. hive_table_partitionkeys
Relationship: Storage Description One table to one storage description. hive_table_storagedesc
Relationship: DDL One table to many DDL entities. hive_table_ddl_queries

Hive Column

Identifier Example Content
typeName hive_column
comment Metadata from Hive from the column description.
name Column name as reported by HMS.
owner Table owner name as reported by HMS.
position This column’s position in the list of columns in a zero-based index.
qualifiedName <database>.<table>.<column>@<clustername>
table Table name. Also modeled as relationship.
type Column data type as reported by HMS.
Relationship: table One table to one or more columns. hive_table_columns
Relationship: inputToProcesses The hive_column_lineage entities that include this column in the input to a transformation. The relationship type is dataset_process_inputs.
Relationship: outputFromProcesses The hive_column_lineage entities that include this column in the output to a transformation. The relationship type is process_dataset_outputs.
Relationship: Table One table to one or more columns. hive_table_columns
Relationship: Partition Key Column One table to one or more columns that are partition keys. hive_table_partitionkeys

Hive Column Lineage

Identifier Example Content
typeName hive_column_lineage
dependencyType The type of relationship between the input and output columns; one of SIMPLE, EXPRESSION, or SCRIPT.
name <database>.<table>@<clustername>:<generated ID>:<output_column>
inputs

List of 0 or more hive_column entities that contributed to the output columns. This is a legacy model component: the more current model uses a relationship attribute.

outputs This is a legacy model component: the more current model uses a relationship attribute.
qualifiedName Same as name.
query Name of the hive_process entity that produced this lineage. This is a legacy model component: the more current model uses a relationship attribute.
Relationship: Process Name of the hive_process entity that produced this lineage. hive_process_column_lineage
Relationship: inputToProcesses List of 0 or more hive_column entities that contributed to the output columns.
Relationship: outputFromProcesses List of 0 or more hive_column entities that were produced in the process.

Hive Storage Description

Identifier Example Content
typeName hive_storagedesc
compressed Metadata from Hive indicating whether the table is stored compressed.
inputFormat Metadata from Hive indicating the storage input format.
outputFormat Metadata from Hive indicating the storage output format.
parameters Additional metadata from Hive in the form of key-value pairs.
qualifiedName <database>.<table>@<clustername>_storage
serdeInfo Metadata from Hive indicating the serialization/deserialization implementation used to write/read table data.
sortCols Metadata from Hive listing the column or columns used to sort the table data.
storedAsSubDirectories Metadata from Hive indicating whether a skewed table uses the list bucketing feature, which creates subdirectories for skewed values.
numBuckets Metadata from Hive indicating the number of buckets for bucketed tables. Non-bucketed tables are indicated by -1.
table The table that this storage description holds data for. Also represented as a relationship.
Relationship: table The table that this storage description holds data for.