Overview
Learn how to use GREG to merge and reconcile multiple CSV datasets into a single golden record. Includes setup steps, key concepts, trust order, and common issues.
GREG is a data reconciliation tool that allows users to compare and merge multiple datasets into a single, unified output—commonly referred to as a golden record. It is designed for working with asset inventories, user records, or any structured data where multiple sources may report on the same entities with variations.
GREG simplifies the process of identifying discrepancies, aligning values, and producing a clean, consolidated dataset ready for reporting, license analysis, or operational use.
What GREG Does
-
Ingests multiple .csv files with differing structures or field names
-
Aligns and compares records based on defined keys
-
Highlights value conflicts and merges data using a defined trust order
-
Produces a golden record that retains the most reliable and prioritized values across sources
Use Cases
-
IT asset reconciliation
Compare data from Active Directory, CMDB, and discovery tools to identify gaps or inconsistencies. -
License audit preparation
Align multiple data sources before submitting data for software audits. -
Data cleanup
Validate and standardize device or user information across business units or systems. -
Multi-source verification
Cross-check different datasets (e.g. vendor exports, internal systems, spreadsheets) for reliability.
How to Use GREG
1. Upload CSV Files
-
GREG accepts .csv files only.
-
You can upload two or more datasets for comparison.
-
Files can contain different field structures and column headers.
2. Map Columns
-
GREG will automatically detect and suggest column mappings.
-
Manual drag-and-drop is available to align fields across datasets.
-
All mapped fields must relate to the same data concept (e.g., "hostname" from each dataset).
3. Define Primary or Composite Keys
To accurately match records across datasets, GREG requires a key:
-
Primary Key: A single field (e.g., hostname, serial_number, or employee_id) that uniquely identifies each record.
-
Composite Key: A combination of two or more fields (e.g., hostname + domain) used when no single field alone is unique across all records.
These keys ensure that GREG compares the right records across sources when building the golden record.
4. Set Trust Order
When the same field (e.g., operating_system) has different values across datasets, GREG uses the trust order to decide which value should be used in the final output.
-
You assign a priority order to each dataset during setup as per the colours.
-
For each conflicting value, GREG selects the value from the highest-priority dataset where the field is populated.
-
This allows you to determine which source is most authoritative per field.
5. Compare Sources
The Compare function in GREG allows you to review how values for each record differ across the uploaded datasets. It highlights:
- Whether a conflict exists for a specific field
- The original values from each source
- Which value was selected as the final, resolved value based on the trust order
This view helps you validate how GREG has reconciled the data, identify inconsistencies, and ensure the merged output is accurate before exporting the golden record.
6. Generate the Golden Record
After mapping and review, GREG compiles a merged dataset.
-
Matching values are unified.
-
Conflicting values are resolved using the trust order.
-
The output can be reviewed in-app and downloaded for further use.
FAQs & Troubleshooting
❓ Can I upload Excel or other file formats?
No, GREG only supports .csv files.
❓ What if my datasets don’t share a common unique field?
Use a composite key by selecting two or more fields that, together, uniquely identify records across datasets.
❓ How many datasets can I compare?
There’s no hard limit—you can upload and map multiple datasets in a single session.
❓ Can I change mappings after uploading?
Yes. You can revisit and modify mappings, primary/composite keys, and trust order settings before generating the final output.
❓ Are deleted records or unmapped fields included in the final output?
Only mapped fields and records matched via keys are included in the final golden record. Unmatched records will still appear but may be flagged for review.