Schema Reflection¶

TiberoDialect implements a broad set of SQLAlchemy reflection hooks against Tibero system views.

Schema normalization¶

Effective schema is derived by _effective_schema(connection, schema).
If schema is provided, uppercase is used.
If omitted, current schema is read with:

SELECT SYS_CONTEXT('USERENV', 'CURRENT_SCHEMA') FROM DUAL

The dialect also enables SQLAlchemy name normalization (requires_name_normalize = True).

Table and view reflection¶

`get_table_names(connection, schema=None, **kw)`¶

With schema: reads all_tables
Without schema: reads user_tables

`get_view_names(connection, schema=None, **kw)`¶

Reads all_views for the effective schema.

`get_view_definition(connection, view_name, schema=None, **kw)`¶

Returns text from all_views for one view, or None.

`get_table_comment(connection, table_name, schema=None, **kw)`¶

Returns { "text": <comment-or-None> } from all_tab_comments.

Column reflection¶

`get_columns(connection, table_name, schema=None, **kw)`¶

Reads:

column_name
data_type
data_length
data_precision
data_scale
nullable
data_default

Type resolution is delegated to _resolve_column_type(...) using ischema_names.

Example:

from sqlalchemy import create_engine, inspect

engine = create_engine("tibero://tibero:password@localhost:8629/TESTDB")
insp = inspect(engine)
print(insp.get_columns("USERS"))

Constraint and index reflection¶

Primary key¶

get_pk_constraint(connection, table_name, schema=None, **kw) returns:

{"name": "PK_USERS", "constrained_columns": ["ID"]}

Foreign keys¶

get_foreign_keys(connection, table_name, schema=None, **kw) groups rows by constraint name and returns constrained/referred columns in order.

Indexes¶

get_indexes(connection, table_name, schema=None, **kw) returns dictionaries with:

name
column_names
unique

Unique constraints¶

get_unique_constraints(connection, table_name, schema=None, **kw) returns named column groups.

Check constraints¶

get_check_constraints(connection, table_name, schema=None, **kw) returns non-null check SQL text.

Sequence reflection¶

`has_sequence(connection, sequence_name, schema=None, **kw)`¶

Implemented via all_sequences count lookup.

`get_sequence_names()` and `get_sequence_range()`¶

These methods are not overridden in this dialect module. If your workflow depends on them, use explicit SQL against Tibero metadata or test SQLAlchemy default behavior in your version.

Additional object existence checks¶

has_table(connection, table_name, schema=None, **kw) checks both tables and views.
has_index(connection, table_name, index_name, schema=None) checks all_indexes.

Reflection flow¶

mermaid flowchart TD i[Inspector call] --> s[_effective_schema] s --> q[Run metadata query on ALL_* or USER_* views] q --> r[_row_get extracts row values] r --> t[_resolve_column_type / group constraints] t --> out[SQLAlchemy reflection dictionaries]

Uppercase object matching

Reflection SQL uppercases table/view/sequence names before filtering.

Unknown type fallback

Unmapped DATA_TYPE values become NULLTYPE, which can impact autogeneration and comparisons.