django1/django/db/backends/oracle/introspection.py

336 lines
14 KiB
Python

from collections import namedtuple
import cx_Oracle
from django.db import models
from django.db.backends.base.introspection import (
BaseDatabaseIntrospection, FieldInfo as BaseFieldInfo, TableInfo,
)
from django.utils.functional import cached_property
FieldInfo = namedtuple('FieldInfo', BaseFieldInfo._fields + ('is_autofield', 'is_json'))
class DatabaseIntrospection(BaseDatabaseIntrospection):
cache_bust_counter = 1
# Maps type objects to Django Field types.
@cached_property
def data_types_reverse(self):
if self.connection.cx_oracle_version < (8,):
return {
cx_Oracle.BLOB: 'BinaryField',
cx_Oracle.CLOB: 'TextField',
cx_Oracle.DATETIME: 'DateField',
cx_Oracle.FIXED_CHAR: 'CharField',
cx_Oracle.FIXED_NCHAR: 'CharField',
cx_Oracle.INTERVAL: 'DurationField',
cx_Oracle.NATIVE_FLOAT: 'FloatField',
cx_Oracle.NCHAR: 'CharField',
cx_Oracle.NCLOB: 'TextField',
cx_Oracle.NUMBER: 'DecimalField',
cx_Oracle.STRING: 'CharField',
cx_Oracle.TIMESTAMP: 'DateTimeField',
}
else:
return {
cx_Oracle.DB_TYPE_DATE: 'DateField',
cx_Oracle.DB_TYPE_BINARY_DOUBLE: 'FloatField',
cx_Oracle.DB_TYPE_BLOB: 'BinaryField',
cx_Oracle.DB_TYPE_CHAR: 'CharField',
cx_Oracle.DB_TYPE_CLOB: 'TextField',
cx_Oracle.DB_TYPE_INTERVAL_DS: 'DurationField',
cx_Oracle.DB_TYPE_NCHAR: 'CharField',
cx_Oracle.DB_TYPE_NCLOB: 'TextField',
cx_Oracle.DB_TYPE_NVARCHAR: 'CharField',
cx_Oracle.DB_TYPE_NUMBER: 'DecimalField',
cx_Oracle.DB_TYPE_TIMESTAMP: 'DateTimeField',
cx_Oracle.DB_TYPE_VARCHAR: 'CharField',
}
def get_field_type(self, data_type, description):
if data_type == cx_Oracle.NUMBER:
precision, scale = description[4:6]
if scale == 0:
if precision > 11:
return 'BigAutoField' if description.is_autofield else 'BigIntegerField'
elif 1 < precision < 6 and description.is_autofield:
return 'SmallAutoField'
elif precision == 1:
return 'BooleanField'
elif description.is_autofield:
return 'AutoField'
else:
return 'IntegerField'
elif scale == -127:
return 'FloatField'
elif data_type == cx_Oracle.NCLOB and description.is_json:
return 'JSONField'
return super().get_field_type(data_type, description)
def get_table_list(self, cursor):
"""Return a list of table and view names in the current database."""
cursor.execute("""
SELECT table_name, 't'
FROM user_tables
WHERE
NOT EXISTS (
SELECT 1
FROM user_mviews
WHERE user_mviews.mview_name = user_tables.table_name
)
UNION ALL
SELECT view_name, 'v' FROM user_views
UNION ALL
SELECT mview_name, 'v' FROM user_mviews
""")
return [TableInfo(self.identifier_converter(row[0]), row[1]) for row in cursor.fetchall()]
def get_table_description(self, cursor, table_name):
"""
Return a description of the table with the DB-API cursor.description
interface.
"""
# user_tab_columns gives data default for columns
cursor.execute("""
SELECT
user_tab_cols.column_name,
user_tab_cols.data_default,
CASE
WHEN user_tab_cols.collation = user_tables.default_collation
THEN NULL
ELSE user_tab_cols.collation
END collation,
CASE
WHEN user_tab_cols.char_used IS NULL
THEN user_tab_cols.data_length
ELSE user_tab_cols.char_length
END as internal_size,
CASE
WHEN user_tab_cols.identity_column = 'YES' THEN 1
ELSE 0
END as is_autofield,
CASE
WHEN EXISTS (
SELECT 1
FROM user_json_columns
WHERE
user_json_columns.table_name = user_tab_cols.table_name AND
user_json_columns.column_name = user_tab_cols.column_name
)
THEN 1
ELSE 0
END as is_json
FROM user_tab_cols
LEFT OUTER JOIN
user_tables ON user_tables.table_name = user_tab_cols.table_name
WHERE user_tab_cols.table_name = UPPER(%s)
""", [table_name])
field_map = {
column: (internal_size, default if default != 'NULL' else None, collation, is_autofield, is_json)
for column, default, collation, internal_size, is_autofield, is_json in cursor.fetchall()
}
self.cache_bust_counter += 1
cursor.execute("SELECT * FROM {} WHERE ROWNUM < 2 AND {} > 0".format(
self.connection.ops.quote_name(table_name),
self.cache_bust_counter))
description = []
for desc in cursor.description:
name = desc[0]
internal_size, default, collation, is_autofield, is_json = field_map[name]
name = name % {} # cx_Oracle, for some reason, doubles percent signs.
description.append(FieldInfo(
self.identifier_converter(name), *desc[1:3], internal_size, desc[4] or 0,
desc[5] or 0, *desc[6:], default, collation, is_autofield, is_json,
))
return description
def identifier_converter(self, name):
"""Identifier comparison is case insensitive under Oracle."""
return name.lower()
def get_sequences(self, cursor, table_name, table_fields=()):
cursor.execute("""
SELECT
user_tab_identity_cols.sequence_name,
user_tab_identity_cols.column_name
FROM
user_tab_identity_cols,
user_constraints,
user_cons_columns cols
WHERE
user_constraints.constraint_name = cols.constraint_name
AND user_constraints.table_name = user_tab_identity_cols.table_name
AND cols.column_name = user_tab_identity_cols.column_name
AND user_constraints.constraint_type = 'P'
AND user_tab_identity_cols.table_name = UPPER(%s)
""", [table_name])
# Oracle allows only one identity column per table.
row = cursor.fetchone()
if row:
return [{
'name': self.identifier_converter(row[0]),
'table': self.identifier_converter(table_name),
'column': self.identifier_converter(row[1]),
}]
# To keep backward compatibility for AutoFields that aren't Oracle
# identity columns.
for f in table_fields:
if isinstance(f, models.AutoField):
return [{'table': table_name, 'column': f.column}]
return []
def get_relations(self, cursor, table_name):
"""
Return a dictionary of {field_name: (field_name_other_table, other_table)}
representing all relationships to the given table.
"""
table_name = table_name.upper()
cursor.execute("""
SELECT ca.column_name, cb.table_name, cb.column_name
FROM user_constraints, USER_CONS_COLUMNS ca, USER_CONS_COLUMNS cb
WHERE user_constraints.table_name = %s AND
user_constraints.constraint_name = ca.constraint_name AND
user_constraints.r_constraint_name = cb.constraint_name AND
ca.position = cb.position""", [table_name])
return {
self.identifier_converter(field_name): (
self.identifier_converter(rel_field_name),
self.identifier_converter(rel_table_name),
) for field_name, rel_table_name, rel_field_name in cursor.fetchall()
}
def get_key_columns(self, cursor, table_name):
cursor.execute("""
SELECT ccol.column_name, rcol.table_name AS referenced_table, rcol.column_name AS referenced_column
FROM user_constraints c
JOIN user_cons_columns ccol
ON ccol.constraint_name = c.constraint_name
JOIN user_cons_columns rcol
ON rcol.constraint_name = c.r_constraint_name
WHERE c.table_name = %s AND c.constraint_type = 'R'""", [table_name.upper()])
return [
tuple(self.identifier_converter(cell) for cell in row)
for row in cursor.fetchall()
]
def get_primary_key_column(self, cursor, table_name):
cursor.execute("""
SELECT
cols.column_name
FROM
user_constraints,
user_cons_columns cols
WHERE
user_constraints.constraint_name = cols.constraint_name AND
user_constraints.constraint_type = 'P' AND
user_constraints.table_name = UPPER(%s) AND
cols.position = 1
""", [table_name])
row = cursor.fetchone()
return self.identifier_converter(row[0]) if row else None
def get_constraints(self, cursor, table_name):
"""
Retrieve any constraints or keys (unique, pk, fk, check, index) across
one or more columns.
"""
constraints = {}
# Loop over the constraints, getting PKs, uniques, and checks
cursor.execute("""
SELECT
user_constraints.constraint_name,
LISTAGG(LOWER(cols.column_name), ',') WITHIN GROUP (ORDER BY cols.position),
CASE user_constraints.constraint_type
WHEN 'P' THEN 1
ELSE 0
END AS is_primary_key,
CASE
WHEN user_constraints.constraint_type IN ('P', 'U') THEN 1
ELSE 0
END AS is_unique,
CASE user_constraints.constraint_type
WHEN 'C' THEN 1
ELSE 0
END AS is_check_constraint
FROM
user_constraints
LEFT OUTER JOIN
user_cons_columns cols ON user_constraints.constraint_name = cols.constraint_name
WHERE
user_constraints.constraint_type = ANY('P', 'U', 'C')
AND user_constraints.table_name = UPPER(%s)
GROUP BY user_constraints.constraint_name, user_constraints.constraint_type
""", [table_name])
for constraint, columns, pk, unique, check in cursor.fetchall():
constraint = self.identifier_converter(constraint)
constraints[constraint] = {
'columns': columns.split(','),
'primary_key': pk,
'unique': unique,
'foreign_key': None,
'check': check,
'index': unique, # All uniques come with an index
}
# Foreign key constraints
cursor.execute("""
SELECT
cons.constraint_name,
LISTAGG(LOWER(cols.column_name), ',') WITHIN GROUP (ORDER BY cols.position),
LOWER(rcols.table_name),
LOWER(rcols.column_name)
FROM
user_constraints cons
INNER JOIN
user_cons_columns rcols ON rcols.constraint_name = cons.r_constraint_name AND rcols.position = 1
LEFT OUTER JOIN
user_cons_columns cols ON cons.constraint_name = cols.constraint_name
WHERE
cons.constraint_type = 'R' AND
cons.table_name = UPPER(%s)
GROUP BY cons.constraint_name, rcols.table_name, rcols.column_name
""", [table_name])
for constraint, columns, other_table, other_column in cursor.fetchall():
constraint = self.identifier_converter(constraint)
constraints[constraint] = {
'primary_key': False,
'unique': False,
'foreign_key': (other_table, other_column),
'check': False,
'index': False,
'columns': columns.split(','),
}
# Now get indexes
cursor.execute("""
SELECT
ind.index_name,
LOWER(ind.index_type),
LISTAGG(LOWER(cols.column_name), ',') WITHIN GROUP (ORDER BY cols.column_position),
LISTAGG(cols.descend, ',') WITHIN GROUP (ORDER BY cols.column_position)
FROM
user_ind_columns cols, user_indexes ind
WHERE
cols.table_name = UPPER(%s) AND
NOT EXISTS (
SELECT 1
FROM user_constraints cons
WHERE ind.index_name = cons.index_name
) AND cols.index_name = ind.index_name
GROUP BY ind.index_name, ind.index_type
""", [table_name])
for constraint, type_, columns, orders in cursor.fetchall():
constraint = self.identifier_converter(constraint)
constraints[constraint] = {
'primary_key': False,
'unique': False,
'foreign_key': None,
'check': False,
'index': True,
'type': 'idx' if type_ == 'normal' else type_,
'columns': columns.split(','),
'orders': orders.split(','),
}
return constraints