deb-python-oslo.db/oslo_db/sqlalchemy/utils.py
Doug Hellmann c0a43732e8 Update Oslo imports to remove namespace package
Change-Id: I4ec9b2a310471e4e07867073e9577731ac34027d
Blueprint: drop-namespace-packages
2015-01-21 18:23:56 -05:00

1015 lines
36 KiB
Python

# Copyright 2010 United States Government as represented by the
# Administrator of the National Aeronautics and Space Administration.
# Copyright 2010-2011 OpenStack Foundation.
# Copyright 2012 Justin Santa Barbara
# All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
import collections
import logging
import re
from oslo_utils import timeutils
import six
import sqlalchemy
from sqlalchemy import Boolean
from sqlalchemy import CheckConstraint
from sqlalchemy import Column
from sqlalchemy.engine import Connectable
from sqlalchemy.engine import reflection
from sqlalchemy.engine import url as sa_url
from sqlalchemy.ext.compiler import compiles
from sqlalchemy import func
from sqlalchemy import Index
from sqlalchemy import inspect
from sqlalchemy import Integer
from sqlalchemy import MetaData
from sqlalchemy.sql.expression import literal_column
from sqlalchemy.sql.expression import UpdateBase
from sqlalchemy.sql import text
from sqlalchemy import String
from sqlalchemy import Table
from sqlalchemy.types import NullType
from oslo_db import exception
from oslo_db._i18n import _, _LI, _LW
from oslo_db.sqlalchemy import models
# NOTE(ochuprykov): Add references for backwards compatibility
InvalidSortKey = exception.InvalidSortKey
ColumnError = exception.ColumnError
LOG = logging.getLogger(__name__)
_DBURL_REGEX = re.compile(r"[^:]+://([^:]+):([^@]+)@.+")
def get_callable_name(function):
# TODO(harlowja): Replace this once
# it is possible to use https://review.openstack.org/#/c/122495/ which is
# a more complete and expansive module that does a similar thing...
try:
method_self = six.get_method_self(function)
except AttributeError:
method_self = None
if method_self is not None:
if isinstance(method_self, six.class_types):
im_class = method_self
else:
im_class = type(method_self)
try:
parts = (im_class.__module__, function.__qualname__)
except AttributeError:
parts = (im_class.__module__, im_class.__name__, function.__name__)
else:
try:
parts = (function.__module__, function.__qualname__)
except AttributeError:
parts = (function.__module__, function.__name__)
return '.'.join(parts)
def sanitize_db_url(url):
match = _DBURL_REGEX.match(url)
if match:
return '%s****:****%s' % (url[:match.start(1)], url[match.end(2):])
return url
# copy from glance/db/sqlalchemy/api.py
def paginate_query(query, model, limit, sort_keys, marker=None,
sort_dir=None, sort_dirs=None):
"""Returns a query with sorting / pagination criteria added.
Pagination works by requiring a unique sort_key, specified by sort_keys.
(If sort_keys is not unique, then we risk looping through values.)
We use the last row in the previous page as the 'marker' for pagination.
So we must return values that follow the passed marker in the order.
With a single-valued sort_key, this would be easy: sort_key > X.
With a compound-values sort_key, (k1, k2, k3) we must do this to repeat
the lexicographical ordering:
(k1 > X1) or (k1 == X1 && k2 > X2) or (k1 == X1 && k2 == X2 && k3 > X3)
We also have to cope with different sort_directions.
Typically, the id of the last row is used as the client-facing pagination
marker, then the actual marker object must be fetched from the db and
passed in to us as marker.
:param query: the query object to which we should add paging/sorting
:param model: the ORM model class
:param limit: maximum number of items to return
:param sort_keys: array of attributes by which results should be sorted
:param marker: the last item of the previous page; we returns the next
results after this value.
:param sort_dir: direction in which results should be sorted (asc, desc)
:param sort_dirs: per-column array of sort_dirs, corresponding to sort_keys
:rtype: sqlalchemy.orm.query.Query
:return: The query with sorting/pagination added.
"""
if 'id' not in sort_keys:
# TODO(justinsb): If this ever gives a false-positive, check
# the actual primary key, rather than assuming its id
LOG.warning(_LW('Id not in sort_keys; is sort_keys unique?'))
assert(not (sort_dir and sort_dirs))
# Default the sort direction to ascending
if sort_dirs is None and sort_dir is None:
sort_dir = 'asc'
# Ensure a per-column sort direction
if sort_dirs is None:
sort_dirs = [sort_dir for _sort_key in sort_keys]
assert(len(sort_dirs) == len(sort_keys))
# Add sorting
for current_sort_key, current_sort_dir in zip(sort_keys, sort_dirs):
try:
sort_dir_func = {
'asc': sqlalchemy.asc,
'desc': sqlalchemy.desc,
}[current_sort_dir]
except KeyError:
raise ValueError(_("Unknown sort direction, "
"must be 'desc' or 'asc'"))
try:
sort_key_attr = inspect(model).\
all_orm_descriptors[current_sort_key]
except KeyError:
raise exception.InvalidSortKey()
query = query.order_by(sort_dir_func(sort_key_attr))
# Add pagination
if marker is not None:
marker_values = []
for sort_key in sort_keys:
v = getattr(marker, sort_key)
marker_values.append(v)
# Build up an array of sort criteria as in the docstring
criteria_list = []
for i in range(len(sort_keys)):
crit_attrs = []
for j in range(i):
model_attr = getattr(model, sort_keys[j])
crit_attrs.append((model_attr == marker_values[j]))
model_attr = getattr(model, sort_keys[i])
if sort_dirs[i] == 'desc':
crit_attrs.append((model_attr < marker_values[i]))
else:
crit_attrs.append((model_attr > marker_values[i]))
criteria = sqlalchemy.sql.and_(*crit_attrs)
criteria_list.append(criteria)
f = sqlalchemy.sql.or_(*criteria_list)
query = query.filter(f)
if limit is not None:
query = query.limit(limit)
return query
def _read_deleted_filter(query, db_model, deleted):
if 'deleted' not in db_model.__table__.columns:
raise ValueError(_("There is no `deleted` column in `%s` table. "
"Project doesn't use soft-deleted feature.")
% db_model.__name__)
default_deleted_value = db_model.__table__.c.deleted.default.arg
if deleted:
query = query.filter(db_model.deleted != default_deleted_value)
else:
query = query.filter(db_model.deleted == default_deleted_value)
return query
def _project_filter(query, db_model, project_id):
if 'project_id' not in db_model.__table__.columns:
raise ValueError(_("There is no `project_id` column in `%s` table.")
% db_model.__name__)
if isinstance(project_id, (list, tuple, set)):
query = query.filter(db_model.project_id.in_(project_id))
else:
query = query.filter(db_model.project_id == project_id)
return query
def model_query(model, session, args=None, **kwargs):
"""Query helper for db.sqlalchemy api methods.
This accounts for `deleted` and `project_id` fields.
:param model: Model to query. Must be a subclass of ModelBase.
:type model: models.ModelBase
:param session: The session to use.
:type session: sqlalchemy.orm.session.Session
:param args: Arguments to query. If None - model is used.
:type args: tuple
Keyword arguments:
:keyword project_id: If present, allows filtering by project_id(s).
Can be either a project_id value, or an iterable of
project_id values, or None. If an iterable is passed,
only rows whose project_id column value is on the
`project_id` list will be returned. If None is passed,
only rows which are not bound to any project, will be
returned.
:type project_id: iterable,
model.__table__.columns.project_id.type,
None type
:keyword deleted: If present, allows filtering by deleted field.
If True is passed, only deleted entries will be
returned, if False - only existing entries.
:type deleted: bool
Usage:
.. code-block:: python
from oslo_db.sqlalchemy import utils
def get_instance_by_uuid(uuid):
session = get_session()
with session.begin()
return (utils.model_query(models.Instance, session=session)
.filter(models.Instance.uuid == uuid)
.first())
def get_nodes_stat():
data = (Node.id, Node.cpu, Node.ram, Node.hdd)
session = get_session()
with session.begin()
return utils.model_query(Node, session=session, args=data).all()
Also you can create your own helper, based on ``utils.model_query()``.
For example, it can be useful if you plan to use ``project_id`` and
``deleted`` parameters from project's ``context``
.. code-block:: python
from oslo_db.sqlalchemy import utils
def _model_query(context, model, session=None, args=None,
project_id=None, project_only=False,
read_deleted=None):
# We suppose, that functions ``_get_project_id()`` and
# ``_get_deleted()`` should handle passed parameters and
# context object (for example, decide, if we need to restrict a user
# to query his own entries by project_id or only allow admin to read
# deleted entries). For return values, we expect to get
# ``project_id`` and ``deleted``, which are suitable for the
# ``model_query()`` signature.
kwargs = {}
if project_id is not None:
kwargs['project_id'] = _get_project_id(context, project_id,
project_only)
if read_deleted is not None:
kwargs['deleted'] = _get_deleted_dict(context, read_deleted)
session = session or get_session()
with session.begin():
return utils.model_query(model, session=session,
args=args, **kwargs)
def get_instance_by_uuid(context, uuid):
return (_model_query(context, models.Instance, read_deleted='yes')
.filter(models.Instance.uuid == uuid)
.first())
def get_nodes_data(context, project_id, project_only='allow_none'):
data = (Node.id, Node.cpu, Node.ram, Node.hdd)
return (_model_query(context, Node, args=data, project_id=project_id,
project_only=project_only)
.all())
"""
if not issubclass(model, models.ModelBase):
raise TypeError(_("model should be a subclass of ModelBase"))
query = session.query(model) if not args else session.query(*args)
if 'deleted' in kwargs:
query = _read_deleted_filter(query, model, kwargs['deleted'])
if 'project_id' in kwargs:
query = _project_filter(query, model, kwargs['project_id'])
return query
def get_table(engine, name):
"""Returns an sqlalchemy table dynamically from db.
Needed because the models don't work for us in migrations
as models will be far out of sync with the current data.
.. warning::
Do not use this method when creating ForeignKeys in database migrations
because sqlalchemy needs the same MetaData object to hold information
about the parent table and the reference table in the ForeignKey. This
method uses a unique MetaData object per table object so it won't work
with ForeignKey creation.
"""
metadata = MetaData()
metadata.bind = engine
return Table(name, metadata, autoload=True)
class InsertFromSelect(UpdateBase):
"""Form the base for `INSERT INTO table (SELECT ... )` statement."""
def __init__(self, table, select):
self.table = table
self.select = select
@compiles(InsertFromSelect)
def visit_insert_from_select(element, compiler, **kw):
"""Form the `INSERT INTO table (SELECT ... )` statement."""
return "INSERT INTO %s %s" % (
compiler.process(element.table, asfrom=True),
compiler.process(element.select))
def _get_not_supported_column(col_name_col_instance, column_name):
try:
column = col_name_col_instance[column_name]
except KeyError:
msg = _("Please specify column %s in col_name_col_instance "
"param. It is required because column has unsupported "
"type by SQLite.")
raise exception.ColumnError(msg % column_name)
if not isinstance(column, Column):
msg = _("col_name_col_instance param has wrong type of "
"column instance for column %s It should be instance "
"of sqlalchemy.Column.")
raise exception.ColumnError(msg % column_name)
return column
def drop_old_duplicate_entries_from_table(migrate_engine, table_name,
use_soft_delete, *uc_column_names):
"""Drop all old rows having the same values for columns in uc_columns.
This method drop (or mark ad `deleted` if use_soft_delete is True) old
duplicate rows form table with name `table_name`.
:param migrate_engine: Sqlalchemy engine
:param table_name: Table with duplicates
:param use_soft_delete: If True - values will be marked as `deleted`,
if False - values will be removed from table
:param uc_column_names: Unique constraint columns
"""
meta = MetaData()
meta.bind = migrate_engine
table = Table(table_name, meta, autoload=True)
columns_for_group_by = [table.c[name] for name in uc_column_names]
columns_for_select = [func.max(table.c.id)]
columns_for_select.extend(columns_for_group_by)
duplicated_rows_select = sqlalchemy.sql.select(
columns_for_select, group_by=columns_for_group_by,
having=func.count(table.c.id) > 1)
for row in migrate_engine.execute(duplicated_rows_select).fetchall():
# NOTE(boris-42): Do not remove row that has the biggest ID.
delete_condition = table.c.id != row[0]
is_none = None # workaround for pyflakes
delete_condition &= table.c.deleted_at == is_none
for name in uc_column_names:
delete_condition &= table.c[name] == row[name]
rows_to_delete_select = sqlalchemy.sql.select(
[table.c.id]).where(delete_condition)
for row in migrate_engine.execute(rows_to_delete_select).fetchall():
LOG.info(_LI("Deleting duplicated row with id: %(id)s from table: "
"%(table)s"), dict(id=row[0], table=table_name))
if use_soft_delete:
delete_statement = table.update().\
where(delete_condition).\
values({
'deleted': literal_column('id'),
'updated_at': literal_column('updated_at'),
'deleted_at': timeutils.utcnow()
})
else:
delete_statement = table.delete().where(delete_condition)
migrate_engine.execute(delete_statement)
def _get_default_deleted_value(table):
if isinstance(table.c.id.type, Integer):
return 0
if isinstance(table.c.id.type, String):
return ""
raise exception.ColumnError(_("Unsupported id columns type"))
def _restore_indexes_on_deleted_columns(migrate_engine, table_name, indexes):
table = get_table(migrate_engine, table_name)
insp = reflection.Inspector.from_engine(migrate_engine)
real_indexes = insp.get_indexes(table_name)
existing_index_names = dict(
[(index['name'], index['column_names']) for index in real_indexes])
# NOTE(boris-42): Restore indexes on `deleted` column
for index in indexes:
if 'deleted' not in index['column_names']:
continue
name = index['name']
if name in existing_index_names:
column_names = [table.c[c] for c in existing_index_names[name]]
old_index = Index(name, *column_names, unique=index["unique"])
old_index.drop(migrate_engine)
column_names = [table.c[c] for c in index['column_names']]
new_index = Index(index["name"], *column_names, unique=index["unique"])
new_index.create(migrate_engine)
def change_deleted_column_type_to_boolean(migrate_engine, table_name,
**col_name_col_instance):
if migrate_engine.name == "sqlite":
return _change_deleted_column_type_to_boolean_sqlite(
migrate_engine, table_name, **col_name_col_instance)
insp = reflection.Inspector.from_engine(migrate_engine)
indexes = insp.get_indexes(table_name)
table = get_table(migrate_engine, table_name)
old_deleted = Column('old_deleted', Boolean, default=False)
old_deleted.create(table, populate_default=False)
table.update().\
where(table.c.deleted == table.c.id).\
values(old_deleted=True).\
execute()
table.c.deleted.drop()
table.c.old_deleted.alter(name="deleted")
_restore_indexes_on_deleted_columns(migrate_engine, table_name, indexes)
def _change_deleted_column_type_to_boolean_sqlite(migrate_engine, table_name,
**col_name_col_instance):
insp = reflection.Inspector.from_engine(migrate_engine)
table = get_table(migrate_engine, table_name)
columns = []
for column in table.columns:
column_copy = None
if column.name != "deleted":
if isinstance(column.type, NullType):
column_copy = _get_not_supported_column(col_name_col_instance,
column.name)
else:
column_copy = column.copy()
else:
column_copy = Column('deleted', Boolean, default=0)
columns.append(column_copy)
constraints = [constraint.copy() for constraint in table.constraints]
meta = table.metadata
new_table = Table(table_name + "__tmp__", meta,
*(columns + constraints))
new_table.create()
indexes = []
for index in insp.get_indexes(table_name):
column_names = [new_table.c[c] for c in index['column_names']]
indexes.append(Index(index["name"], *column_names,
unique=index["unique"]))
c_select = []
for c in table.c:
if c.name != "deleted":
c_select.append(c)
else:
c_select.append(table.c.deleted == table.c.id)
ins = InsertFromSelect(new_table, sqlalchemy.sql.select(c_select))
migrate_engine.execute(ins)
table.drop()
for index in indexes:
index.create(migrate_engine)
new_table.rename(table_name)
new_table.update().\
where(new_table.c.deleted == new_table.c.id).\
values(deleted=True).\
execute()
def change_deleted_column_type_to_id_type(migrate_engine, table_name,
**col_name_col_instance):
if migrate_engine.name == "sqlite":
return _change_deleted_column_type_to_id_type_sqlite(
migrate_engine, table_name, **col_name_col_instance)
insp = reflection.Inspector.from_engine(migrate_engine)
indexes = insp.get_indexes(table_name)
table = get_table(migrate_engine, table_name)
new_deleted = Column('new_deleted', table.c.id.type,
default=_get_default_deleted_value(table))
new_deleted.create(table, populate_default=True)
deleted = True # workaround for pyflakes
table.update().\
where(table.c.deleted == deleted).\
values(new_deleted=table.c.id).\
execute()
table.c.deleted.drop()
table.c.new_deleted.alter(name="deleted")
_restore_indexes_on_deleted_columns(migrate_engine, table_name, indexes)
def _change_deleted_column_type_to_id_type_sqlite(migrate_engine, table_name,
**col_name_col_instance):
# NOTE(boris-42): sqlaclhemy-migrate can't drop column with check
# constraints in sqlite DB and our `deleted` column has
# 2 check constraints. So there is only one way to remove
# these constraints:
# 1) Create new table with the same columns, constraints
# and indexes. (except deleted column).
# 2) Copy all data from old to new table.
# 3) Drop old table.
# 4) Rename new table to old table name.
insp = reflection.Inspector.from_engine(migrate_engine)
meta = MetaData(bind=migrate_engine)
table = Table(table_name, meta, autoload=True)
default_deleted_value = _get_default_deleted_value(table)
columns = []
for column in table.columns:
column_copy = None
if column.name != "deleted":
if isinstance(column.type, NullType):
column_copy = _get_not_supported_column(col_name_col_instance,
column.name)
else:
column_copy = column.copy()
else:
column_copy = Column('deleted', table.c.id.type,
default=default_deleted_value)
columns.append(column_copy)
def is_deleted_column_constraint(constraint):
# NOTE(boris-42): There is no other way to check is CheckConstraint
# associated with deleted column.
if not isinstance(constraint, CheckConstraint):
return False
sqltext = str(constraint.sqltext)
# NOTE(I159): in order to omit the CHECK constraint corresponding
# to `deleted` column we have to test these patterns which may
# vary depending on the SQLAlchemy version used.
constraint_markers = (
"deleted in (0, 1)",
"deleted IN (:deleted_1, :deleted_2)",
"deleted IN (:param_1, :param_2)"
)
return any(sqltext.endswith(marker) for marker in constraint_markers)
constraints = []
for constraint in table.constraints:
if not is_deleted_column_constraint(constraint):
constraints.append(constraint.copy())
new_table = Table(table_name + "__tmp__", meta,
*(columns + constraints))
new_table.create()
indexes = []
for index in insp.get_indexes(table_name):
column_names = [new_table.c[c] for c in index['column_names']]
indexes.append(Index(index["name"], *column_names,
unique=index["unique"]))
ins = InsertFromSelect(new_table, table.select())
migrate_engine.execute(ins)
table.drop()
for index in indexes:
index.create(migrate_engine)
new_table.rename(table_name)
deleted = True # workaround for pyflakes
new_table.update().\
where(new_table.c.deleted == deleted).\
values(deleted=new_table.c.id).\
execute()
# NOTE(boris-42): Fix value of deleted column: False -> "" or 0.
deleted = False # workaround for pyflakes
new_table.update().\
where(new_table.c.deleted == deleted).\
values(deleted=default_deleted_value).\
execute()
def get_connect_string(backend, database, user=None, passwd=None,
host='localhost'):
"""Get database connection
Try to get a connection with a very specific set of values, if we get
these then we'll run the tests, otherwise they are skipped
DEPRECATED: this function is deprecated and will be removed from oslo.db
in a few releases. Please use the provisioning system for dealing
with URLs and database provisioning.
"""
args = {'backend': backend,
'user': user,
'passwd': passwd,
'host': host,
'database': database}
if backend == 'sqlite':
template = '%(backend)s:///%(database)s'
else:
template = "%(backend)s://%(user)s:%(passwd)s@%(host)s/%(database)s"
return template % args
def is_backend_avail(backend, database, user=None, passwd=None):
"""Return True if the given backend is available.
DEPRECATED: this function is deprecated and will be removed from oslo.db
in a few releases. Please use the provisioning system to access
databases based on backend availability.
"""
from oslo_db.sqlalchemy import provision
connect_uri = get_connect_string(backend=backend,
database=database,
user=user,
passwd=passwd)
try:
eng = provision.Backend._ensure_backend_available(connect_uri)
eng.dispose()
except exception.BackendNotAvailable:
return False
else:
return True
def get_db_connection_info(conn_pieces):
database = conn_pieces.path.strip('/')
loc_pieces = conn_pieces.netloc.split('@')
host = loc_pieces[1]
auth_pieces = loc_pieces[0].split(':')
user = auth_pieces[0]
password = ""
if len(auth_pieces) > 1:
password = auth_pieces[1].strip()
return (user, password, database, host)
def index_exists(migrate_engine, table_name, index_name):
"""Check if given index exists.
:param migrate_engine: sqlalchemy engine
:param table_name: name of the table
:param index_name: name of the index
"""
inspector = reflection.Inspector.from_engine(migrate_engine)
indexes = inspector.get_indexes(table_name)
index_names = [index['name'] for index in indexes]
return index_name in index_names
def add_index(migrate_engine, table_name, index_name, idx_columns):
"""Create an index for given columns.
:param migrate_engine: sqlalchemy engine
:param table_name: name of the table
:param index_name: name of the index
:param idx_columns: tuple with names of columns that will be indexed
"""
table = get_table(migrate_engine, table_name)
if not index_exists(migrate_engine, table_name, index_name):
index = Index(
index_name, *[getattr(table.c, col) for col in idx_columns]
)
index.create()
else:
raise ValueError("Index '%s' already exists!" % index_name)
def drop_index(migrate_engine, table_name, index_name):
"""Drop index with given name.
:param migrate_engine: sqlalchemy engine
:param table_name: name of the table
:param index_name: name of the index
"""
table = get_table(migrate_engine, table_name)
for index in table.indexes:
if index.name == index_name:
index.drop()
break
else:
raise ValueError("Index '%s' not found!" % index_name)
def change_index_columns(migrate_engine, table_name, index_name, new_columns):
"""Change set of columns that are indexed by given index.
:param migrate_engine: sqlalchemy engine
:param table_name: name of the table
:param index_name: name of the index
:param new_columns: tuple with names of columns that will be indexed
"""
drop_index(migrate_engine, table_name, index_name)
add_index(migrate_engine, table_name, index_name, new_columns)
def column_exists(engine, table_name, column):
"""Check if table has given column.
:param engine: sqlalchemy engine
:param table_name: name of the table
:param column: name of the colmn
"""
t = get_table(engine, table_name)
return column in t.c
class DialectFunctionDispatcher(object):
@classmethod
def dispatch_for_dialect(cls, expr, multiple=False):
"""Provide dialect-specific functionality within distinct functions.
e.g.::
@dispatch_for_dialect("*")
def set_special_option(engine):
pass
@set_special_option.dispatch_for("sqlite")
def set_sqlite_special_option(engine):
return engine.execute("sqlite thing")
@set_special_option.dispatch_for("mysql+mysqldb")
def set_mysqldb_special_option(engine):
return engine.execute("mysqldb thing")
After the above registration, the ``set_special_option()`` function
is now a dispatcher, given a SQLAlchemy ``Engine``, ``Connection``,
URL string, or ``sqlalchemy.engine.URL`` object::
eng = create_engine('...')
result = set_special_option(eng)
The filter system supports two modes, "multiple" and "single".
The default is "single", and requires that one and only one function
match for a given backend. In this mode, the function may also
have a return value, which will be returned by the top level
call.
"multiple" mode, on the other hand, does not support return
arguments, but allows for any number of matching functions, where
each function will be called::
# the initial call sets this up as a "multiple" dispatcher
@dispatch_for_dialect("*", multiple=True)
def set_options(engine):
# set options that apply to *all* engines
@set_options.dispatch_for("postgresql")
def set_postgresql_options(engine):
# set options that apply to all Postgresql engines
@set_options.dispatch_for("postgresql+psycopg2")
def set_postgresql_psycopg2_options(engine):
# set options that apply only to "postgresql+psycopg2"
@set_options.dispatch_for("*+pyodbc")
def set_pyodbc_options(engine):
# set options that apply to all pyodbc backends
Note that in both modes, any number of additional arguments can be
accepted by member functions. For example, to populate a dictionary of
options, it may be passed in::
@dispatch_for_dialect("*", multiple=True)
def set_engine_options(url, opts):
pass
@set_engine_options.dispatch_for("mysql+mysqldb")
def _mysql_set_default_charset_to_utf8(url, opts):
opts.setdefault('charset', 'utf-8')
@set_engine_options.dispatch_for("sqlite")
def _set_sqlite_in_memory_check_same_thread(url, opts):
if url.database in (None, 'memory'):
opts['check_same_thread'] = False
opts = {}
set_engine_options(url, opts)
The driver specifiers are of the form:
``<database | *>[+<driver | *>]``. That is, database name or "*",
followed by an optional ``+`` sign with driver or "*". Omitting
the driver name implies all drivers for that database.
"""
if multiple:
cls = DialectMultiFunctionDispatcher
else:
cls = DialectSingleFunctionDispatcher
return cls().dispatch_for(expr)
_db_plus_driver_reg = re.compile(r'([^+]+?)(?:\+(.+))?$')
def dispatch_for(self, expr):
def decorate(fn):
dbname, driver = self._parse_dispatch(expr)
if fn is self:
fn = fn._last
self._last = fn
self._register(expr, dbname, driver, fn)
return self
return decorate
def _parse_dispatch(self, text):
m = self._db_plus_driver_reg.match(text)
if not m:
raise ValueError("Couldn't parse database[+driver]: %r" % text)
return m.group(1) or '*', m.group(2) or '*'
def __call__(self, *arg, **kw):
target = arg[0]
return self._dispatch_on(
self._url_from_target(target), target, arg, kw)
def _url_from_target(self, target):
if isinstance(target, Connectable):
return target.engine.url
elif isinstance(target, six.string_types):
if "://" not in target:
target_url = sa_url.make_url("%s://" % target)
else:
target_url = sa_url.make_url(target)
return target_url
elif isinstance(target, sa_url.URL):
return target
else:
raise ValueError("Invalid target type: %r" % target)
def dispatch_on_drivername(self, drivername):
"""Return a sub-dispatcher for the given drivername.
This provides a means of calling a different function, such as the
"*" function, for a given target object that normally refers
to a sub-function.
"""
dbname, driver = self._db_plus_driver_reg.match(drivername).group(1, 2)
def go(*arg, **kw):
return self._dispatch_on_db_driver(dbname, "*", arg, kw)
return go
def _dispatch_on(self, url, target, arg, kw):
dbname, driver = self._db_plus_driver_reg.match(
url.drivername).group(1, 2)
if not driver:
driver = url.get_dialect().driver
return self._dispatch_on_db_driver(dbname, driver, arg, kw)
def _invoke_fn(self, fn, arg, kw):
return fn(*arg, **kw)
class DialectSingleFunctionDispatcher(DialectFunctionDispatcher):
def __init__(self):
self.reg = collections.defaultdict(dict)
def _register(self, expr, dbname, driver, fn):
fn_dict = self.reg[dbname]
if driver in fn_dict:
raise TypeError("Multiple functions for expression %r" % expr)
fn_dict[driver] = fn
def _matches(self, dbname, driver):
for db in (dbname, '*'):
subdict = self.reg[db]
for drv in (driver, '*'):
if drv in subdict:
return subdict[drv]
else:
raise ValueError(
"No default function found for driver: %r" %
("%s+%s" % (dbname, driver)))
def _dispatch_on_db_driver(self, dbname, driver, arg, kw):
fn = self._matches(dbname, driver)
return self._invoke_fn(fn, arg, kw)
class DialectMultiFunctionDispatcher(DialectFunctionDispatcher):
def __init__(self):
self.reg = collections.defaultdict(
lambda: collections.defaultdict(list))
def _register(self, expr, dbname, driver, fn):
self.reg[dbname][driver].append(fn)
def _matches(self, dbname, driver):
if driver != '*':
drivers = (driver, '*')
else:
drivers = ('*', )
for db in (dbname, '*'):
subdict = self.reg[db]
for drv in drivers:
for fn in subdict[drv]:
yield fn
def _dispatch_on_db_driver(self, dbname, driver, arg, kw):
for fn in self._matches(dbname, driver):
if self._invoke_fn(fn, arg, kw) is not None:
raise TypeError(
"Return value not allowed for "
"multiple filtered function")
dispatch_for_dialect = DialectFunctionDispatcher.dispatch_for_dialect
def get_non_innodb_tables(connectable, skip_tables=('migrate_version',
'alembic_version')):
"""Get a list of tables which don't use InnoDB storage engine.
:param connectable: a SQLAlchemy Engine or a Connection instance
:param skip_tables: a list of tables which might have a different
storage engine
"""
query_str = """
SELECT table_name
FROM information_schema.tables
WHERE table_schema = :database AND
engine != 'InnoDB'
"""
params = {}
if skip_tables:
params = dict(
('skip_%s' % i, table_name)
for i, table_name in enumerate(skip_tables)
)
placeholders = ', '.join(':' + p for p in params)
query_str += ' AND table_name NOT IN (%s)' % placeholders
params['database'] = connectable.engine.url.database
query = text(query_str)
noninnodb = connectable.execute(query, **params)
return [i[0] for i in noninnodb]