SUMMARY:

• Teradata DBA, Data Warehouse & Performance Architect with expertise in the large scale & high performance physical database design, SQL performance tuning, with achievements of Teradata Uncompromising Quality Award and 20% overall database performance improvement for the Teradata massive parallel processing system.
• For 17 years at the Teradata Company since the first release, I was a senior developer and then Advisory Member in the Database Development as well as a performance specialist in the Professional Service with in - depth knowledge of Teradata database and SQL.
• For 13 years as the Teradata DBA, Data Warehouse & Performance Architect, I have done extensive query performance tuning via EXPLAIN, DBQL, Viewpoint, and optimized physical database design for ETL and BI tools like DataStage and Cognos.

EMPLOYMENT HISTORY:

Teradata Data Warehouse & Performance Architect

Confidential, Huntington Beach, CA

Responsibilities:

• Enhanced ETL, TPT, & Cognos of Enterprise Finance Business Intelligence System.
• Converted 285 SQLs of the DataStage Change-Data-Capture (TPT Export & Load, file comparisons) to 7 Balanced Optimizer SQLs, 4 INSERT-SELECT & 3 DELETE.
• Changed hundreds of DataStage ETL from the Teradata TPT Export and Load API to the INSERT-SELECT.
• Reduced CPU & IO 10-20 times
• Performance enhancement on slow Cognos reports by the optimized partitioned primary index.
• Reduced 99% CPU & IO
• Query optimization by avoiding large number of rows per primary index values.
• Teradata distributes tables horizontally across all AMP processors based on the primary index. Large number of rows per primary index value of a single table join index on the fact table would cause performance problems in joins and data maintenance (insert/delete/update/alter table). The solution is to collect statistics on join columns but dropping such Join Indexes.
• Reduced CPU 10 times and IO & Spool 3-4 times
• Optimized a query that took 70K CPU to remove all rows of the associate group except the min(EFF DATE). For example was kept as the min(EFF DATE) of the associate group and was deleted.

Teradata DBA Specialist, Data Warehouse & Performance Architect

Confidential, Pasadena, CA

Responsibilities:

• Implemented the Data Base Query Log (DBQL) Normalization.
• Performed weekly statistics collection on all 94,000 reporting tables with 212,000 statistics and 50 terabytes space by implementing scripts and macros to retain statistics data in a table in order to instantly identify, collect, and restore stale, lost, and missing statistics of tables on a database, as well as providing estimated row count of all tables without counting day-by-day.
• Optimized ETL staging commands to reduced 90% CPU and 95% IO on large tables.
• Redesigned the job divided staging code on a large table to double the performance by converting 78 queries of 18 staging commands to 18 queries of 6 staging commands.
• This task includes 6 developments. Each loads 3 job-divided files into 3 staging tables.
• When one development deletes, updates, and inserts data in the same reporting table via the Begin/End Transaction staging commands, others are blocked while holding system resources. The elapsed time of the last blocked query would be 6 times longer.
• Four such block-each-other tasks in the Teradata system would take all 20 threads of the ETL Windows console to hold hundred of tasks in the thread queue.
• Further enhancement can load all job-divided files from AIX server into one staging table and uses 1 staging command with 3 queries to the reporting table with no block.
• Identified & resolved the system performance issues via the AMP Work Task (AWT), flow control, and Teradata Active Systems Management (TASM).

Teradata Data Warehouse & Performance Architect

Confidential, Pleasanton, CA

Responsibilities:

• Implemented dual-active data warehouse (data replication with no duplicate process):
• Performance tuning for ad hoc queries and application scripts via EXPLAIN Report, Performance Monitor, Join Index, and Partitioned Primary Index.
• Provides multi-statement inserts with write row hash lock and join indexes to optimize Web applications to have less than a 5 second combined response time when executing the key queries inserts and updates.
• Project architect to implement the Teradata Customer Relationship Management System to provide segmentation, campaign management and statistical modeling and rules-based personalized offer. The marketing users could quickly identify households by cluster, product purchasing history and other behavioral attributes. Offers could be easily created, maintained and tracked to deliver a relevant incentive to each household, while driving key strategic corporate goals.
• Implemented Supply Chain Performance Management System to integrate data from Confidential operational systems and data warehouse with SeeCommerce Demand & Inventory Application to support Retail Promotion Performance and Vendor Scorecard requirements.

Sr. Teradata Consultant

Confidential, Los Angeles, CA

Responsibilities:

• Designed and implemented Teradata ETL to Extract, Transform, and Load thousands of Planogram files via Wal-Mart Retail Link Web, Unix Korn Shell, FTP, FastLoad, FastExport, BTEQ, and SQL scripts with 7,500 lines of code for Warner Bros.
• Developed Teradata ETL to Extract, Transform, and Load MarketNavigator files. It combines demographics, survey results, census and sales data to enable the optimization of home video marketing, promotion and distribution for Warner Bros.
• Designed and implemented the inventory database applications.

Sr. Database Developer

Confidential, El Segundo, CA

Responsibilities:

• Implemented the logical and physical database design and development of Object-Relational Database.
• Prototyped the ANSI SQL-99 Object-Relational Database Data Dictionary tables and Information Schema views. Systematically converted existing dictionary table data from Relational to Object-Relational Database, written in C.
• Enhanced Embedded Structured Query Language to support large object functions.

Sr. Consulting Engineer

Confidential, El Segundo, CA

Responsibilities:

• Provided the performance expertise in the Teradata Professional Service.
• Project architect and leader to redesign and port the Campaign Management database application, a third generation Data Warehouse Customer Relationship Management (CRM), to analyze customer information, to perform backend analysis and to turn the marketing planning processes from product-focused to customer event driven from UNIX to NT via ODBC and Perl. All data Extract, Transform and Load (ETL), BTEQ, FastExport, FastLoad, and MultiLoad, were either rewritten to reduce the row size from 3,000 to 100 bytes via data modeling or replaced by scripted SQL queries to double the performance.
• Performed database backup, restoration, performance tuning and capacity planning.
• Quickly provided a business solution overnight with a complete logical and physical database design and the On Line Analytical Processing development to resolve the last minute change of a business requirement when the Confidential Korea Business Solutions Team installed the CRM.
• Analyzed and implemented the query optimization to meet the pharmacy application performance requirement on the 100 TB Wal-Mart Data Warehouse in one day just before the initial production.
• Designed and implemented to double the fundamental Teradata database design limit to 64K of Data Block, Row, and Field by changing 20,000 lines over 1,000 modules in one-person year to achieve 20% overall system performance improvement and 75% fast in full table scan. Coordinated the integration process of this mass code change into a major Teradata release with strong people & communication skills.
• Implemented the Teradata SQL to meet ANSI SQL by a strong knowledge of SQL.
• Developed Teradata Client Utilities, BTEQ and ITEQ.
• Defined compiler requirements and developed the Microsoft-to-MetaWare Pascal conversion tool to port the Teradata database software to UNIX.
• Developed conversion tools to port the Teradata database software from Pascal to C.
• Analyzed the advanced features of Pascal to enhance the performance of this Teradata System Development Language; designed the code optimization enhancements. Thousands of database modules relied on this bug-free compiler since 1984.