Data Standards¶

The most common quality issue in data modeling is naming inconsistency. If the same concept of "Customer Number" appears as CUST_NO, CUST_NUM, CUSTOMER_NUMBER, and CUSTID across different tables, the cost of identifying relationships and maintaining the system grows exponentially.

ThinkERD fundamentally solves this problem with its Word → Term → Domain three-tier standardization system.

Why Data Standardization Matters¶

Modeling without standards

Table	Column Name	Intent
Customer	`CUST_NO`	Customer Number
Order	`CUSTOMER_NUMBER`	Customer Number
Payment	`CUSTID`	Customer Number
Shipping	`CUST_NUM`	Customer Number

4 different physical names for the same concept. As the system scales, this inconsistency becomes structural technical debt.

With a standardization system, Customer Number always resolves to CUST_NO. Since physical names are auto-generated when modelers enter logical names, the cost of standards compliance approaches zero.

Three-Tier Structure¶

graph TB
    W["🔤 Standard Word"] -->|"compose"| T["📝 Standard Term"]
    D["📐 Standard Domain"] -->|"bind"| T
    T -->|"apply to column"| C["📊 Entity Column"]

    style W fill:#e0f2fe,stroke:#0284c7
    style T fill:#f0fdf4,stroke:#16a34a
    style D fill:#fef3c7,stroke:#d97706
    style C fill:#f3e8ff,stroke:#9333ea

Standard Word¶

The smallest mapping unit between logical and physical abbreviations. The foundation for all naming standards.

Logical Name	Physical Abbr.	Type
Customer	`CUST`	Modifier
Number	`NO`	Suffix
Date	`DT`	Suffix
Amount	`AMT`	Suffix

Standard Term¶

A standardized attribute name composed of words. Ensures logical↔physical consistency for column names.

Logical Composition	Physical Result
Customer + Number	`CUST_NO`
Order + Date	`ORD_DT`
Payment + Amount	`PAY_AMT`

Standard Domain¶

A type standard that defines data type, length, and constraints. Binding a domain to a term automatically determines the column's data type.

Domain Name	Data Type	Usage Example
Code(10)	`VARCHAR(10)`	Category code, status code
Amount	`DECIMAL(18,2)`	Order amount, payment amount
Date	`DATE`	Order date, created date
Name(100)	`VARCHAR(100)`	Customer name, product name

Domain Benefits

When a domain changes, all columns using that domain are updated at once. For example, changing the Amount domain precision from DECIMAL(18,2) to DECIMAL(20,4) automatically updates all amount columns that reference it.

Standard Management Tools¶

Words, Terms & Domains

Registration, composition, and binding methods for the three-tier standard system, with practical usage patterns

Details
Standard Studio

Dedicated workspace for managing large volumes of standard words and terms. Search, filter, and batch-edit capabilities

Details
AI Naming

Uses a local LLM to auto-generate standard physical names from logical names. Learns abbreviation rules and prohibited words for high-accuracy results

Details