Intelligence · Module
Data Fabric
Bring ERP, spreadsheet and warehouse data in, with quality checks and lineage
Data Fabric is planned to ingest data from outside One3 through incremental pipelines with quality rules, purpose tags and full lineage. It joins external records to core entities by cross-reference and never writes into app-owned tables.
When important data lives elsewhere
Many businesses keep key numbers in an ERP, Tally, a warehouse or a set of MIS spreadsheets. Data Fabric, planned for W3, brings that data into One3 safely so it can sit beside app data in dashboards, BI Studio and agents.
Sources
Planned sources run through Integration Hub connections: PostgreSQL, MySQL, SQL Server, CSV, XLSX and Parquet files from object storage, Google Sheets, and business apps prioritised in discovery such as Zoho, Shopify and SAP. Tally is planned with Finance in W4. A spreadsheet upload wizard detects headers and types and supports recurring upload slots for monthly MIS files. For warehouse users, read connectors for Snowflake, BigQuery, Databricks and Microsoft Fabric are planned so data can be read in place. India Stack sources such as GST and e-invoice data via a GSP and Account Aggregator data through a licensed route are planned for W4 (each licence and access route to be verified).
Pipelines and transformations
Pipelines load incrementally with checkpoints, idempotent loads, schema inference and a dead-letter queue. Schedules, retries, backfills, run history and cancellation are durable. Transformations are versioned SQL models with tests and documentation, and a visual preparation flow compiles no-code steps into those models for review as a diff.
Quality and contracts
Quality rules cover not-null, uniqueness, accepted values, ranges, referential integrity to core entities, freshness and reconciliation against a source total. Each dataset gets a quality score. Publication is atomic: consumers keep reading the last good version while a new version fails quality. Schema drift is checked against data contracts, and breaking changes pause dependants and alert owners. Observability monitors learn normal ranges for volume, freshness and null rates.
Privacy by default
PII is detected and classified on ingest, masking policies apply, and datasets containing personal data cannot be published without consent-purpose tags. Pinned-region tenants cannot configure staging or destinations outside the region. Lineage runs from source to pipeline to dataset to metric to dashboard, alert and agent. Sync volume is metered and each pipeline has a cost view.
Features
What Data Fabric includes
Source connections
PostgreSQL, MySQL, SQL Server, object storage files, Google Sheets and prioritised business apps via Integration Hub.
Incremental pipelines
Checkpoints, idempotent loads, schema inference and a dead-letter queue.
Versioned SQL transformations
Tests, run order and documentation for each model.
Visual data preparation
Filter, join, union, aggregate, pivot, dedupe and lookup steps compiled to reviewed SQL.
Dataset registry
Schema, owner, classification, purpose tags, freshness target, row counts and version.
Data-quality rules and scores
Not-null, uniqueness, ranges, referential integrity, freshness and reconciliation checks.
Atomic publication
Consumers read the last good version while a failing version is held back.
Schema drift and data contracts
Breaking changes pause dependants and alert owners.
PII detection and purpose tags
Classification on ingest, masking policies, and no publishing of personal data without purpose tags.
End-to-end lineage
Source to pipeline to dataset to metric to dashboard, alert and agent.
Spreadsheet upload wizard
Header detection, type inference and recurring upload slots for MIS files.
Warehouse coexistence
Read connectors for Snowflake, BigQuery, Databricks and Microsoft Fabric (verify API terms).
AI functions in preparation
Classify, extract, summarise, translate or mask as a step, with preview, budget and cost.
Metering and cost view
Sync volume meter and per-pipeline cost.
Workflow
Source to certified use
- 1
Connect a source through Integration Hub
- 2
Generate or write the pipeline and mapping
- 3
Run quality rules and PII classification
- 4
Add purpose tags and publish the dataset version
- 5
Join to core entities by cross-reference in BI Studio
- 6
Use in certified metrics with lineage shown
AI
AI skills and Digital Employees
Pipeline generator
Drafts a pipeline, mapping and quality rules from a description and sample; publishing is human.
Data-quality agent
Detects rule failures and drift and proposes fixes.
Schema drift explanation
Summarises contract diffs for owners.
External record mapping
Maps external records to core entities using match rules, never overwriting core fields.
Digital Employees that use this module
They act through the same governed Actions as people; risky steps wait for approval.
Frequently asked questions
Does Data Fabric change data in my One3 apps?
No. It never writes into app-owned tables. Links to core entities are cross-references.
Can I bring my monthly Excel MIS?
Yes. A spreadsheet upload wizard with recurring upload slots is planned for W3.
Is GST or Account Aggregator data supported?
They are planned for W4 through licensed routes. Each licence and access route must be verified first.
See Data Fabric in your workflow
Tell us about your team and the outcomes you want. We will walk you through the apps, the Digital Employees and the plan for your industry.