td-acf

name	td-acf
description	Auto-correlation analysis for time series dependency and pattern detection

Teradata Auto-Correlation Function

Skill Name	Teradata Auto-Correlation Function
Description	Auto-correlation analysis for time series dependency and pattern detection
Category	UAF Time Series
Function	TD_ACF
Framework	Teradata Unbounded Array Framework (UAF)

Core Capabilities

• Advanced UAF implementation with EXECUTE FUNCTION INTO ART syntax
• Scalable time series analysis for millions of products or billions of IoT sensors
• SERIES_SPEC-based input with TABLE_NAME, ROW_AXIS, SERIES_ID, PAYLOAD
• Production-ready SQL generation with proper UAF syntax
• Comprehensive error handling and data validation
• Business-focused interpretation of analytical results
• Integration with UAF pipeline workflows

UAF Function Syntax

TD_ACF uses the EXECUTE FUNCTION INTO VOLATILE ART(result_table) pattern:

EXECUTE FUNCTION COLUMNS(OUT_{value_col} AS {value_col}) INTO VOLATILE ART(acf_results)
TD_ACF(
    SERIES_SPEC(
        TABLE_NAME({input_table}),
        SERIES_ID({series_id_col}),
        ROW_AXIS(TIMECODE({time_column})),
        PAYLOAD(FIELDS({value_col}), CONTENT(REAL))
    ),
    FUNC_PARAMS(MAXLAGS({max_lags}), FUNC_TYPE(0), QSTAT(1), DEMEAN(1), ALPHA(0.05))
);
SELECT * FROM acf_results;

Key Parameters for TD_ACF

• MAXLAGS: Maximum number of lags to compute (integer)
• FUNC_TYPE: 0 for standard ACF
• QSTAT: 1 to compute Ljung-Box Q-statistics
• DEMEAN: 1 to subtract mean before computing ACF
• ALPHA: Significance level for confidence intervals (e.g., 0.05)

How to Use This Skill

1. Provide Your Time Series Data:

   "Analyze time series table: database.sensor_data with timestamp column and value columns"
   

2. The Skill Will:

   • Analyze temporal structure and sampling frequency
   • Identify optimal UAF function parameters
   • Generate complete TD_ACF workflow using EXECUTE FUNCTION INTO ART syntax
   • Provide performance optimization recommendations

Input Requirements

Data Requirements

• Time series table: Teradata table with temporal data
• Timestamp column: Time/date column for temporal indexing (used in ROW_AXIS TIMECODE)
• Value columns: Numeric columns for analysis (used in PAYLOAD FIELDS)
• Series ID column: Column identifying distinct series (used in SERIES_ID)
• Regular sampling: Consistent time intervals (recommended)

Technical Requirements

• Teradata Vantage with UAF (Unbounded Array Framework) enabled
• UAF License: Access to time series and signal processing functions
• Database permissions: EXECUTE FUNCTION, SELECT on working database
• Function access: TD_ACF

Output

TD_ACF results are stored in a volatile ART table. Use SELECT * FROM result_table to read them.

Result Columns

• Lag number
• Auto-correlation coefficient at each lag
• Q-statistic (if QSTAT=1)
• P-value for Q-statistic
• Confidence interval bounds (based on ALPHA)

Scripts Included

Core UAF Scripts

• uaf_data_preparation.sql: UAF-specific data preparation
• td_acf_workflow.sql: Complete TD_ACF implementation with EXECUTE FUNCTION syntax
• td_acf_workflow_template.sql: Parameterized template
• uaf_table_analysis.sql: Time series structure analysis
• parameter_optimization.sql: Function parameter tuning

Integration Scripts

• uaf_pipeline_template.sql: Multi-function UAF workflows using ART chaining
• performance_monitoring.sql: UAF execution monitoring
• result_interpretation.sql: Output analysis and visualization

UAF Best Practices

• Use EXECUTE FUNCTION INTO VOLATILE ART(...) for all UAF function calls
• Chain UAF functions by referencing ART result tables in subsequent SERIES_SPEC TABLE_NAME
• Always specify SERIES_ID even for single-series data
• Use ROW_AXIS(TIMECODE(...)) for time-indexed series
• Specify PAYLOAD with FIELDS and CONTENT(REAL) for numeric data

---

This skill provides production-ready UAF time series analytics using Teradata's Unbounded Array Framework TD_ACF with correct EXECUTE FUNCTION INTO ART syntax.