Generate JPA entity classes from Oracle CREATE TABLE DDL statements. Use when the user provides a CREATE TABLE definition and wants a Hibernate entity. Handles Oracle types (VARCHAR2, NUMBER, CLOB, CHAR, DATE, TIMESTAMP, BLOB), schema-qualified tables, and follows the hibertable11 repo conventions.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
Generate JPA entity classes from Oracle CREATE TABLE DDL statements. Use when the user provides a CREATE TABLE definition and wants a Hibernate entity. Handles Oracle types (VARCHAR2, NUMBER, CLOB, CHAR, DATE, TIMESTAMP, BLOB), schema-qualified tables, and follows the hibertable11 repo conventions.
user_invocable
true
Entity Generator Skill
Generate JPA entity classes from Oracle CREATE TABLE DDL following hibertable11 conventions.
Input
User provides a CREATE TABLE statement. Example:
CREATE TABLE RATOR.WF_ACTION_CONFIG
(
UUID CHAR(36CHAR) NOT NULL,
ACTION VARCHAR2(256CHAR),
CONFIG CLOB,
CONFIG_FORMAT VARCHAR2(256CHAR),
ID NUMBER
)
Generation Rules
1. Package Placement
Determine package from table name prefix or schema:
Table starts with OI_ → dk.secondbrand.hibertable11.domain.entity.oister, class prefixed Oi
Table starts with FLEXII_ → dk.secondbrand.hibertable11.domain.entity.flexii, class prefixed Flexii
Otherwise →
dk.secondbrand.hibertable11.domain.entity.shared
If the table name has a known prefix like WF_, DK2_, GENERIC_ — keep it in shared unless user specifies.
Ask the user if placement is ambiguous.
2. Class Name Derivation
Convert UPPER_SNAKE_CASE table name to PascalCase:
Remove schema prefix (e.g., RATOR.)
Remove brand prefix (OI_, FLEXII_) — it becomes the class prefix
WF_ACTION_CONFIG → WfActionConfig
OI_ACCOUNT → OiAccount
FLEXII_USER → FlexiiUser
PROVISIONING_LOG → ProvisioningLog
3. Class Structure (exact annotation order)
package dk.secondbrand.hibertable11.domain.entity.<subpackage>;
import dk.secondbrand.hibertable11.utils.NextOidGenerator;
import jakarta.persistence.*;
import lombok.*;
// Add java.time imports only if date/timestamp columns exist// Add java.math.BigDecimal only if NUMBER with scale > 0@Entity@Table(name = "TABLE_NAME")// Add schema = "SCHEMA" if schema-qualified@Getter@Setter@ToString// Add exclude = {"field"} for @Lob and relationship fields@NoArgsConstructor@AllArgsConstructorpublicclassClassName {
@Id@NextOidGenerator@Column(name = "ID", nullable = false)private Long id;
// ... fields ...
}
4. Oracle → Java Type Mapping
Oracle Type
Java Type
@Column Attributes
NUMBER (no precision)
Long
—
NUMBER(p) where p ≤ 9
Integer
—
NUMBER(p) where p > 9, scale=0
Long
—
NUMBER(p,s) where s > 0
BigDecimal
precision = p, scale = s
VARCHAR2(n CHAR) or VARCHAR2(n)
String
length = n
CHAR(n CHAR) or CHAR(n)
String
length = n
CLOB
String
Add @Lob on separate line above @Column
BLOB
byte[]
Add @Lob + @Basic(fetch = FetchType.LAZY)
DATE
LocalDateTime
—
TIMESTAMP
LocalDateTime
—
TIMESTAMP WITH TIME ZONE
LocalDateTime
—
RAW(16)
String
length = 32 (hex representation)
5. Field Name Derivation
Convert UPPER_SNAKE_CASE column name to camelCase:
STATUS_ID → statusId
CREATE_DATE → createDate
UUID → uuid
CONFIG_FORMAT → configFormat
6. ID Column Handling
If table has an ID NUMBER column → standard @Id @NextOidGenerator @Column(name = "ID", nullable = false) private Long id;
If table has NO ID column but has a UUID CHAR(36) → use UUID as primary key:
Add the field name to @ToString(exclude = {"config"}) to avoid lazy-load issues.
10. Relationship Resolution for FK Columns
For every column ending in _ID that is NOT the primary key, you MUST search the existing codebase to check if a matching entity already exists.
Lookup procedure:
Take the column name, strip the _ID suffix → candidate table name (e.g., WORKFLOW_ID → WORKFLOW)
Also try with common prefixes: OI_, FLEXII_, WF_, GENERIC_ (e.g., WF_WORKFLOW)
Search for existing entities: use Grep/Glob to find @Table(name = "WORKFLOW" or @Table(name = "WF_WORKFLOW" etc. across src/main/java/**/entity/**/*.java
Also check if the column name itself matches a known table (e.g., PROVISIONING_TASK_ID → look for table PROVISIONING_TASK)
If a matching entity IS found:
Generate the full relationship following repo conventions — raw FK Long + read-only @ManyToOne:
Then remaining fields in the order they appear in the CREATE TABLE (excluding ID which was moved to top)
Multiple Tables
If user provides multiple CREATE TABLE statements, generate one entity per table. Ask about shared base classes if tables have identical column subsets.
Inheritance Detection
If two tables (e.g., OI_MAIL_MESSAGE and FLEXII_MAIL_MESSAGE) share most columns:
Suggest a @MappedSuperclass base in shared/
Concrete entities in oister/ and flexii/ extending it
Use @SuperBuilder + @NoArgsConstructor instead of @AllArgsConstructor
