Technical lineage for Tableau

When you ingest Tableau metadata in Data Catalog, a technical lineage for Tableau Data Attribute assets is automatically created.

Permissions

If you have a Data Catalog global role with the Catalog and Technical lineage global permissions, you can see the technical lineage of Tableau assets by clicking on the Technical lineage tab on the asset page of any of the following asset types:

  • Table
  • Column
  • Tableau Data Attribute
  • Tableau Worksheet

Technical lineage graph

The technical lineage graph shows relations of the type "Data Element sources / targets Data Element" between Tableau assets and other data objects in the data flow, for example between a Column asset and a Tableau Data Attribute asset. These relations are created during the Tableau ingestion process as a result of automatic stitching.

Note If you use a Tableau <source ID> configuration file and don’t specify a value for the relevant collibraSystemName property, the designation “UNDEFINED” will be shown in the technical lineage.

Note  If you use custom SQL that is not supported by the Tableau metadata API, the technical lineage might not be complete. For complete information, see the Tableau documentation on Tableau Catalog support for custom SQL and Tableau Lineage and custom SQL connections.

Example

The following technical lineage shows how data flows from a PostgreSQL data source to Tableau. It shows relations of the type "Data Element sources / targets Data Element" between the Column assets of the database and Tableau Data Attribute assets in Tableau. For example, Column asset L_RETURNFLAG has a relation of the type "Data Element sources / targets Data Element" to the Tableau Data Attribute assets Quantity and Adjusted Quantity.

UUIDs in the Tableau technical lineage

Collibra Data Lineage uses unique full names to create a technical lineage and stitch objects within the technical lineage. Full names in Collibra are constructed in accordance with the hierarchy of data objects in Tableau, for example:

  • Server > Site > Project > Workbook > Worksheet > Field
  • Server > Site > Project > Workbook > Data Model > Column

In Collibra, every node in this hierarchy must have a unique name. However, in Tableau, the names of data objects do not have to be unique. As such, if Tableau data objects in a technical lineage hierarchy have the same full name, Collibra Data Lineage adds the UUIDs of the corresponding assets to the names in the technical lineage, to maintain uniqueness.

In the following example image, the names of the assets Priority, Opened and Active in the technical lineage have been appended with their UUIDs.

Note 
  • UUIDs are not added to the names of the assets themselves; they are only added to the names of the data objects in the technical lineage.
  • The UUID is always part of the full name of an asset, regardless of whether or not it is a duplicate.

How to resolve UUIDs in names in a technical lineage

To keep Collibra Data Lineage from adding UUIDs to the names of the data objects in a technical lineage, ensure that the names of all fields and columns in Tableau are unique.

Generally, Tableau doesn't allow you to create two fields or columns with the same name. However, hierarchy fields and non-hierarchy fields can have the same name. Duplication of names can also happen if:

  • A Tableau worksheet is using two different data sources that have columns with the same name.
  • You create a virtual connection that contains multiple data sources that have columns with the same name.
  • There are multiple data sources in Tableau with the same name.

Sources tab page

The Sources tab page shows, for each Tableau data source and Tableau Worksheet, the transformation and calculation rules that the Collibra Data Lineage service analyzed and processed, and the results of the analysis. It also shows the TECHLIN VIEW query definitions, based on custom SQL queries.

If a parameter is used in a Tableau worksheet, it is shown in the worksheet source code, for example:

PARAMETERS: 'parameter1'.

If a parameter is used in a calculation rule, it is also shown under the Tableau data source for data sources in the calculation rule, for example:

CALCULATION RULE: '[List price]/[parameter1]'

The success rate of the analysis indicates how complete the technical lineage is. There are a few limitations that prevent the Collibra Data Lineage service from processing all Tableau metadata.

Important The Collibra Data Lineage service might not be able to process all complex Tableau metadata. This means that the success rate of a Tableau ingestion might not be 100%.

Error codes

The Errors summary represents a summary of all errors per Tableau site. Currently, processing stops after the first error is identified during the processing of the site. In a future version of Collibra, processing will continue after the first error.

Warning codes

Warning codes indicate:

  • Issues that might affect the technical lineage, but do not stop the processing.
  • Issues that you can resolve.

Element Description
ID The warning ID number.
Name

The name of the warning. Possible values are:

  • Empty name
  • Field relation not found
  • Parent project not found
  • Parent workbook not found
  • Parent database not found
  • Datasource not found
  • Worksheet not found
  • REST datasource not found
  • Not found in remote fields
  • Multiple datasources
  • Query parsing error
  • Invalid Collibra system names
  • Invalid hostname mapping
Status code

The status label. The value is always WARNING.
Status description

Identifies a grouping of warnings. Warnings of the same type (meaning they have the same group name and name) are grouped together in "parts" of up to 100 warnings.

Example 

In this example, there are 250 Configuration > Invalid Collibra system names warnings, grouped into parts 1, 2 and 3:
Group name

The type, or category, of warning. Possible values are:

  • Configuration
  • Mismatched ID
  • Missing content

The following table shows the complete set of warning codes, by group and name.

Group name Name Description

Missing content

Empty name

Raised during the processing of databases, tables, columns, worksheets and dashboards.

Contains the following lines:

  • Database with id DATABASE_ID is skipped
  • Table with id TABLE_ID is skipped
  • Column with id COLUMN_ID is skipped
  • Worksheet with id WORKSHEET_ID is skipped
  • Dashboard with id DASHBOARD_ID is skipped

Indicates that the name property of a database, table, column, worksheet or dashboard, which has a specified value for the id property, has a null value or it is empty:

Example for a database:

{
  "data": {
    "databasesConnection": {
      "nodes": [
        {
          "id": "DATABASE_ID",
          "name": null,
           ...

Note The name property is considered empty if the value is null or if it is empty.

Missing content

Parent database not found

Raised during the processing of tables.

Contains the line: Table with id TABLE_ID is skipped

Indicates that the parent database for a table with TABLE_ID was not found in the previously processed databases.

Possible causes:

  • The database property is not present in the JSON file.
  • The database property is empty: "database": {}.
  • The DATABASE_ID is not present for the id property.
{
  "data": {
    "tablesConnection": {
      "nodes": [
        {        
          "id": "TABLE_ID",
          "database": {
             "id": "DATABASE_ID"
          }
          ...

Missing content

Parent project not found

Raised during the processing of projects, workbooks and published data sources.

Contains the following lines:

  • Workbook with id WORKBOOK_ID is skipped
  • Published datasource with id DATASOURCE_ID is skipped
  • Project with id PROJECT_ID has unreachable parent project

Indicates that the parent project of a project, workbook, or published data source was not found in the previously processed projects.

Possible causes:

  • The project property is not present in the JSON file.
  • The project property is empty: "project": {}.
  • The PROJECT_ID is not present for the id property.

Example for a workbook:

{
  "workbooks": {
    "workbook": [
      {
        "project": {
           "id": "PROJECT_ID"
        },
        "id": "WORKBOOK_ID",
        ...

Example for a published datasource:

To identify the PROJECT_ID, first find the DATASOURCE_LUID of the published data source, as returned by the metadata API:

{
  "data": {
    "datasourcesConnection": {
      "nodes": [
        {
          "__typename": "PublishedDatasource",
          "id": "DATASOURCE_ID",
          "luid": "DATASOURCE_LUID"
          ...

Then, in the data returned by the REST API, reference the DATASOURCE_LUID to identify the PROJECT_ID of the data source.:

{
  "datasources": {
    "datasource": [
      {
        "id": "DATASOURCE_LUID",
        "project": {
          "id": "PROJECT_ID",
          ...

Example for a project:

PARENT_PROJECT_ID is not found:

{
  "projects": {
    "project": [
      {
        "id": "PROJECT_ID",
        "parentProjectId": "PARENT_PROJECT_ID",
        ...

Project is not skipped in this case. The new parent project is created with name Unknown project PARENT_PROJECT_ID.

Missing content

Mismatched ID

Parent workbook not found

Raised during processing of worksheets, dashboards, REST-only views and embedded data sources.

Contains the following lines:

  • Worksheet with id WORKSHEET_ID is skipped
  • Dashboard with id DASHBOARD_ID is skipped
  • View with id VIEW_ID is skipped (rest only)
  • Embedded data source with id DATASOURCE_ID is skipped

Indicates that the parent workbook of a worksheet, dashboard or view with a specified ID was not found in the previously processed workbooks.

Possible causes:

  • The workbook property is not present in the JSON file.
  • The workbook property is empty: "workbook": {}.
  • WORKBOOK_ID is not present for the luid property.
  • mismatched ID issue.

Example for a worksheet:

{
  "data": {
    "sheetsConnection": {
      "nodes": [
        {
          "id": "WORKSHEET_ID",                    
          "workbook": {
            "luid": "WORKBOOK_ID"
          ...

Example for a dashboard:

{
  "data": {
    "dashboardsConnection": {
      "nodes": [
        {
          "id": "DASHBOARD_ID",                    
          "workbook": {
            "luid": "WORKBOOK_ID"
          ...

Example for an embedded data source:

{
  "data": {
    "dashboardsConnection": {
      "nodes": [
        {
          "id": "DASHBOARD_ID",                    
          "workbook": {
            "luid": "WORKBOOK_ID"
          ...

Note Use the luid property, not the id property, to find a workbook by ID.

MIssing content

Worksheet not found

Raised during the processing of dashboards.

Contains the line: Worksheet with id WORKSHEET_ID is skipped for dashboard with id DASHBOARD_ID

Indicates that a worksheet with a given ID was not found in the previously processed worksheets.

{
  "data": {
    "dashboardsConnection": {
      "nodes": [
        {
          "id": "DASHBOARD_ID",                    
          "sheets": [
            {
              "id": "WORKSHEET_ID"
            },
          ...

Possible cause: WORKSHEET_ID is not present for the id property.

Mismatched ID

REST datasource not found

Raised during the processing of published data sources.

Contains the line: Published datasource with id DATASOURCE_ID is skipped

Indicates that a data source with DATASOURCE_ID could not be matched with the DATASOURCE_LUID returned by the REST API, resulting in a mismatched ID.

{
  "data": {
    "datasourcesConnection": {
      "nodes": [
        {
          "__typename": "PublishedDatasource",
          "id": "DATASOURCE_ID",
          "luid": "DATASOURCE_LUID"
          ...

During processing, information returned by the metadata API and the REST API is combined. Collibra Data Lineage then looks to the DATASOURCE_LUID property in the REST metadata to identify the correct project ID, which is lacking from the information returned by the metadata API.

This only applies to published data sources, as embedded data sources are assigned to workbooks, not projects.

Missing content

Datasource not found

Raised during the processing of embedded data sources.

Contains the line: Embedded datasource with id EMBEDDED_DATASOURCE_ID references non existing published datasource with id PUBLISHED_DATASOURCE_ID

Indicates that an embedded data source with EMBEDDED_DATASOURCE_ID references a published data source with PUBLISHED_DATASOURCE_ID, which was not found in the previously processed data sources.

{
  "data": {
    "datasourcesConnection": {
      "nodes": [
        {
          "__typename": "EmbeddedDatasource",
          "id": "EMBEDDED_DATASOURCE_ID",
          "upstreamDatasources": [
            {
              "id": "PUBLISHED_DATASOURCE_ID",
              ...

Possible cause: PUBLISHED_DATASOURCE_ID is not present for the id property.

Missing content

Field relation not found

Raised during the processing of data source fields.

Contains the lines:

  • Referenced field with id FIELD_ID is skipped
  • Report field with id FIELD_ID is skipped

Indicates that a field with a given FIELD_ID was not found in remote fields, which is needed to create relations.

{
  "data": {
    "datasourcesConnection": {
      "nodes": [
        {
          "id": "DATASOURCE_ID",
          "fieldsConnection": {
            "nodes": [
              {
                "__typename": "DatasourceField",
                "remoteField": {
                  "id": "FIELD_ID"
                ...

Possible cause: An embedded datasource has a calculated field that is not mapped to any published data source field.

This can occur:

  • During the processing of referenced fields. In this case, the relation between the two Tableau Data Attributes cannot be created.
  • During the processing of report fields. In this case, the relation between the Tableau Data Attribute and the Tableau Data Worksheet cannot be created.

Missing content

Multiple datasources

Raised during the processing of custom SQL queries.

Contains the line: Custom sql query with id QUERY_ID contains columns of NUMBER_OF_DATASOURCES datasources. Found best datasource: DATASOURCE_ID

Indicates that a query with QUERY_ID has matched multiple data sources. Only one data source can be used: datasource with DATASOURCE_ID.

The warning is caused by the fact that there is no direct relation between the query and the data source. The algorithm tries to find the best data source, based on a comparison of the list of query columns and the data source columns.

To verify this, do the following:

  1. Find the query with QUERY_ID and the columns (see COLUMN_ID) in the table JSON data:
    {
  "data": {
    "tablesConnection": {
      "nodes": [
        {
          "__typename": "CustomSQLTable",
          "id": "QUERY_ID",
          "columnsConnection": {
            "nodes": [
              {
                "id": "COLUMN_ID",
                ...
  2. Find the data source with DATASOURCE_ID in the data source JSON data. It should contain all of the columns ( see COLUMN_ID) that are used by the query:
    {
  "data": {
    "datasourcesConnection": {
      "nodes": [
        {
          "id": "DATASOURCE_ID",
          "fieldsConnection": {
            "nodes": [
              {
                "upstreamColumnsConnection": {
                  "nodes": [
                    {
                      "id": "COLUMN_ID"
                    ...

The data source found for this query (meaning DATASOURCE_ID) might not be the right one for the TECHLIN VIEW definition. In this case, the data source DATASOURCE_ID might have the wrong relations between the Tableau Data Attribute asset and the Column asset.

MIssing content

Datasource not found

Raised during the processing of custom SQL queries.

Contains the line: Custom sql query with id QUERY_ID is skipped

Indicates that query with QUERY_ID contains columns that are not referenced by any data source fields, so the data source can’t be assigned to the query.

{
  "data": {
    "tablesConnection": {
      "nodes": [
        {
          "__typename": "CustomSQLTable",
          "id": "QUERY_ID",
          "columnsConnection": {
            "nodes": [
              {
                "id": "COLUMN_ID",
                ...

Missing content

Query parsing error

Raised during the processing of custom SQL queries.

Contains the line: Error parsing query with id QUERY_ID, error: ERROR

Indicates that there is an issue when deriving column names from a query for a custom SQL with QUERY_ID.

{
  "data": {
    "tablesConnection": {
      "nodes": [
        {
          "__typename": "CustomSQLTable",
          "id": "QUERY_ID",
          "query": "QUERY"

Custom SQL is still processed as TECHLIN VIEW with no columns.

Configuration

Invalid Collibra system names

Raised during the processing of the collibraSystemNames section in the <source ID> configuration file.

Contains the lines:

  • Collibra system name not found for database with hostname "DB_HOSTNAME"
  • Collibra system name not found for file with path "FILE_PATH"
  • Collibra system name not found for connector with url "CONNECTION_URL"
  • Collibra system name not found for cloud file with name "CLOUD FILE PATH"

Configuration

Invalid hostname mapping

Raised during the processing of the hostnameMapping section the <source ID> configuration file.

Contains the line: Collibra system name not found for database "DB_NAME" host "HOST_NAME" and schema "SCHEMA_NAME"