# Incremental Model Strategy Selector

> Selects and configures optimal incremental model strategies

- Skill: `a5c-ai/incremental-model-strategy-selector` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add a5c-ai/incremental-model-strategy-selector`
- Raw SKILL.md: https://api.skillmd.com/api/skills/a5c-ai/incremental-model-strategy-selector/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Data & Analytics
- Author: a5c-ai (https://skillmd.com/u/a5c-ai)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/a5c-ai/incremental-model-strategy-selector

---


# Incremental Model Strategy Selector

## Overview

Selects and configures optimal incremental model strategies. This skill optimizes data transformation efficiency through proper incremental processing patterns.

## Capabilities

- Incremental strategy selection (append, merge, delete+insert)
- Partition pruning optimization
- Unique key configuration
- On_schema_change handling
- Full refresh scheduling
- Lookback window optimization
- Late-arriving data handling

## Input Schema

```json
{
  "modelCharacteristics": {
    "sourceType": "string",
    "updatePattern": "append|update|delete",
    "volumeGB": "number",
    "updateFrequency": "string"
  },
  "platform": "snowflake|bigquery|redshift",
  "existingModel": "object"
}
```

## Output Schema

```json
{
  "strategy": "append|merge|delete+insert",
  "config": "object",
  "partitionStrategy": "object",
  "refreshSchedule": "object",
  "dbtConfig": "object"
}
```

## Target Processes

- Incremental Model Setup
- dbt Model Development
- Pipeline Migration

## Usage Guidelines

1. Analyze source data update patterns
2. Measure data volume and update frequency
3. Select strategy based on characteristics
4. Configure appropriate lookback windows

## Best Practices

- Use append for insert-only sources
- Use merge for sources with updates
- Configure partition pruning for large tables
- Schedule periodic full refreshes for data correction
- Handle late-arriving data with appropriate lookback