Source reports change over time. Fields get added, columns get renamed, and report structures evolve. Scoop intelligently handles these changes, automatically adapting your datasets without breaking existing analyses or requiring manual intervention.

How Scoop Handles Report Changes

When you load a new report, Scoop performs intelligent matching:

New Report Arrives
       ↓
┌─────────────────────────────┐
│ 1. Analyze column structure │
│ 2. Detect unique key        │
│ 3. Compare to existing data │
└─────────────────────────────┘
       ↓
   ┌───┴───┐
   │Match? │
   └───┬───┘
    ↓     ↓
  Yes     No
   ↓       ↓
Add to   Create new
existing table or
table    adjust schema

The Matching Process

Step 1: Identify Columns

Scoop examines the incoming report to identify:

• Column names and positions
• Data types (text, number, date)
• Which columns contain unique values

Step 2: Detect Unique Key

A unique key is a column that:

• Uniquely identifies each row (like Opportunity ID)
• Appears exactly once per record
• Remains consistent across report versions

When Scoop finds a unique key, it can confidently match records across schema changes.

Step 3: Match to Existing Tables

Types of Changes Scoop Handles

Adding New Columns

Most common scenario — You add a field to your source report.

Before: [ID, Name, Amount, Stage]
After:  [ID, Name, Amount, Stage, Region, Owner]

What happens: Scoop adds the new columns to the existing table. Historical data will have NULL values for the new columns; new data will have values.

Removing Columns

Scenario: You remove a field you no longer need.

Before: [ID, Name, Amount, Stage, OldField]
After:  [ID, Name, Amount, Stage]

What happens: The column remains in the table (for historical data) but new records won't have values for it.

Renaming Columns

Scenario: Column header name changes.

Before: [ID, Customer_Name, Amount]
After:  [ID, CustomerName, Amount]

What happens: Scoop treats the renamed column as a new column and the old column as removed. To preserve continuity, consider using Calculated Columns to map the old name to the new name.

Reordering Columns

Scenario: Columns appear in a different order.

What happens: Scoop matches by column name, not position. Column order changes don't affect matching.

When Matching Fails

Sometimes Scoop creates a new table instead of adding to an existing one. This happens when the schema difference is significant enough that automatic adaptation isn't safe.

Why a New Table Gets Created

Comparing Tables

When you see multiple tables in your dataset, use the Compare Tables feature:

1. Open your dataset
2. Click the Tables tab
3. Click Compare Tables

The comparison view shows both table schemas side-by-side:

Look for:

• Missing columns — Removed in one version
• Extra columns — Added in one version
• Type differences — Column changed from text to number
• Name variations — Similar names with different spelling

Merging Tables

If two tables should be one dataset:

1. Identify which table has the correct structure
2. Delete the incorrect report loads (see below)
3. Reload reports with consistent structure
4. Reprocess the dataset

Deleting Bad Reports

Occasionally a malformed report gets loaded. You can remove it without affecting other data.

How to Delete a Report

1. Open the dataset
2. Go to Scoop History tab
3. Find the problematic report load
4. Click the trash can icon
5. Confirm deletion
6. Reprocess the dataset

The deleted report is removed from Scoop's history. When you reprocess, that data won't be included.

When to Delete Reports

What Gets Deleted

• The specific report load entry
• Data from that load in the processed dataset (after reprocessing)
• NOT: Other report loads, calculated columns, or analyses

Best Practices

Preparing for Schema Changes

Before making changes to source reports:

1. Document the change — Know what's being added/removed
2. Consider timing — Make changes at the start of a period if possible
3. Verify unique keys — Ensure ID columns remain consistent
4. Test first — Load a sample before full production load

Maintaining Report Consistency

Handling Major Changes

For significant report restructuring:

1. Create a new dataset — Don't force incompatible schemas together
2. Blend with history — Use Blending to combine old and new
3. Map columns — Use calculated columns to align naming
4. Archive old dataset — Keep for historical reference

Troubleshooting

Multiple Tables Appearing

Symptom: Dataset shows 2+ tables when you expected one.

Diagnosis:

1. Compare tables to identify differences
2. Check for schema changes in source
3. Verify unique key consistency

Resolution:

• Delete incorrect loads and reprocess, OR
• Create new dataset with consistent schema

Columns Not Appearing

Symptom: Added a column to source, but it's not in Scoop.

Diagnosis:

1. Check if new load went to a different table
2. Verify column name matches expectations
3. Confirm report was actually reloaded

Resolution:

• Ensure new report loads match existing table structure
• Reprocess dataset to see new columns

Historical Data Issues

Symptom: Old data missing after schema change.

Diagnosis: New table may have been created for new schema.

Resolution:

• Use Compare Tables to understand the split
• Consider blending tables if they should be together
• Accept split if schemas are truly incompatible

Related Topics

• Snapshot Datasets - Tracking changes over time
• Blending Datasets - Combine multiple sources
• Adding Calculated Columns - Map column names
• Reprocessing Data - Rebuild dataset from history