Lineage Glossary

Alation Cloud Service Applies to Alation Cloud Service instances of Alation

Customer Managed Applies to customer-managed instances of Alation

auto-generated lineage

Lineage type as defined by its origin. Auto-generated lineage is automatically generated by the system through query log ingestion (QLI), metadata extraction (MDE), and Compose object resolution pipelines that analyze SQL query history to determine lineage in ingested metadata. Auto-generation of lineage is a default feature, but a source admin can disable it for specific sources in the source settings. If auto-generated lineage is disabled, users can add lineage information manually or upload it with the Lineage API.

badge

Visual indicator on a lineage node that identifies it as temporary, external, or deprecated.

classic layout view

A legacy view of lineage charts that offers a basic set of features to explore lineage. We recommend using the compound layout view instead.

column-level lineage

Lineage type as defined by its granularity. Column-level lineage is presented at the column level for an RDBMS data source, BI field level for a BI source, and file attribute level for a file system. Depending on the source, column-level lineage may require additional enablement on the Alation instance. Specific connector documentation outlines the need for configuring column-level lineage for a connector.

compound layout view

A lineage visualization that groups related lineage nodes into clusters by categories, such as sources, to help you navigate very complex data relationships with multiple interconnected source and target objects and isolate specific areas of interest. The compound layout view provides a sequential display of the data journey from source to target, as well as multiple interactive capabilities, enabling you to adjust the size of the chart, filter out elements, and visualize additional metadata layers, such as deprecations, important notes, and data quality information.

cross-source lineage

Lineage type as defined by source coverage. Cross-source lineage tracks the data flow through multiple data sources within Alation, illustrating how data objects in one catalog source affect objects in another.

dataflow

An object type that stores information about data transformations. On a lineage chart, dataflow objects are represented as separate elements with lineage links entering and exiting, indicating the flow of data through a transformation. Alation’s lineage platform captures various types of dataflows, including stored procedures, ETL/ELT transformations, and SQL queries. Dataflow objects store the code of the transformation from ingested metadata, facilitating lineage investigations.

dataflow filtering

A type of interaction on the classic layout view of lineage charts where you can filter out dataflow elements, leaving only lineage nodes on the chart. This helps declutter the chart when dealing with extensive lineage data. When the dataflow filter is applied, temporary tables connected to the dataflows are also excluded from the chart. The details of the filtered dataflows are organized into a list view, positioned alongside the lineage chart, to provide a clear, consumable format and serve as a side-by-side reference.

default lineage view

The initial lineage chart that has not been customized or modified by a user through interacting with the chart elements. This default state occurs when (a) an Admin hasn’t modified the chart using lineage chart customization options in Admin Settings > Customize Catalog, or (b) a user hasn’t customized the chart by expanding or collapsing vertical or horizontal lineage.

direct lineage extraction

Lineage data obtained directly from system tables of certain database systems (like Databricks Unity Catalog or SAP ECC) rather than inferred from query history through query log ingestion (QLI). Direct lineage extraction is applicable in systems where lineage data is explicitly stored in tables and accessible through metadata extraction (MDE). If a system supports direct lineage extraction, it is generally recommended to disable auto-generated lineage to ensure that only one consistent lineage generation pipeline is used, aligned with the lineage metadata in the database system.

distance

Number of lineage nodes shown upstream and downstream relative to the object from which an Impact Analysis report is accessed. Distance defines the extent of lineage information represented in the report, allowing users to analyze the data flow over the specified range. Also see horizontal lineage.

downstream lineage

Lineage that is located downstream of a specific data object, capturing the flow of data as it moves away from this object as the source object to target objects.

edge

The same as lineage link.

external lineage object

Objects that do not match any existing object type in Alation’s catalog, shown with an external (EXT) badge on a lineage chart.

horizontal lineage

Refers to the breadth of the lineage chart, specifically the number of upstream and downstream nodes visible to users. Also see Distance.

Impact Analysis

Type of the Impact Analysis report that aims at examining the changes within the data ecosystem to evaluate how changes in one area might affect connected downstream areas. This report helps in assessing the impact from such changes as infrastructure migrations or product deployments.

Impact Analysis report

Tabular report that can be generated using lineage data to understand the potential impact of a change within a system on the downstream relationships among data objects. These reports help identify which objects would be affected by changes in the data environment. Impact analysis reports are accessible from lineage charts and the catalog page of the source.

lineage

Information about the origin of data and its movement through data pipelines of an organization’s data ecosystem to destinations where it’s ultimately consumed by users. Lineage documents how target data objects are created from source data objects, as well as transformations that modify data objects as the data flows from sources to targets. In Alation, lineage information is visually represented as a lineage chart on the Lineage tab of data, BI, or file system source catalog pages.

lineage chart

Visual representation of lineage information where data objects are shown as lineage nodes, transformations as dataflow objects, and connections between them are illustrated with arrows (edges, or links) indicating the direction of data flow. Lineage charts offer multiple interactive and analytical features allowing users to explore lineage, such as metadata layering and dataflow filtering. Lineage charts are available for catalog sources that support lineage. Also see compound layout view and classic layout view.

The connection between elements on a lineage chart, illustrating the flow of data and represented as an arrow pointing from an upstream object to a downstream object. A more technical term would be a lineage edge.

lineage path

In manual lineage, denotes a lineage link from an upstream object, or source object, to a dataflow object or from a dataflow object to a downstream object, or target object.

lineage resolution

The process where Alation identifies and matches objects in ingested query history data to data objects in Alation data and BI sources, enhancing the accuracy of lineage charts.

lineage sub-path

In manual lineage, refers to the more granular lineage paths that occur between column objects, nested under table-level or view-level paths.

lineage type

Classification of lineage based on various criteria:

lineage views

Various visual representations of lineage information, such as the compound layout view or the classic layout view.

manual lineage

Lineage type as defined by its origin. Manual lineage is a capability within the Alation catalog that allows users to manually create lineage charts, as opposed to having them automatically generated by the system based on ingested lineage data from metadata extraction (MDE), query log ingestion (QLI), and Compose.

metadata layer

A visual layer on the lineage chart that allows users to add or remove specific metadata, such as trust flag information. This feature is available in the compound layout view.

node

An element on lineage charts that represents a data object, such as a table or view.

node truncation

A feature in large lineage charts where nodes are collapsed under a group node labeled More nodes (classic layout view) or a collapsible panel (compound layout view) to simplify the initial view, with options to expand and explore more details as needed.

single-source lineage

Lineage type as defined by source coverage. Single-source lineage spans across objects within a single source, showing, for example, how tables or views in an RDBMS data source are derived from other tables within the same data source, or how BI reports are created from BI data sources base or BI server.

source object

An upstream lineage object when viewed from the current object being analyzed on the chart. Source objects provide data to the object you’re viewing.

table-level lineage

Lineage type as defined by its granularity. Table-level lineage is visualized at the level of tables or views, which is the default setting.

target object

A downstream lineage object as viewed from the object being analyzed on the lineage chart and originating from this object, with the current object serving as the source object.

temporary lineage object

Objects that Alation cannot find in the catalog (or cannot resolve to an existing object), marked with a temporary (TMP) badge on the lineage chart.

Upstream Audit

A type of the Impact Analysis report that offers an examination of the upstream data sources and inputs that contribute to a specific data object. This audit helps in assessing the origin and quality of data.

upstream lineage

Lineage that is located upstream of a specific data object, capturing the flow of data as it moves toward this object as a target from source objects.

vertical lineage

Refers to the depth of lineage charts in terms of the number of nodes displayed vertically, and the number of columns within a table initially visible before you click to expand further.