I'd like to create a temporary table in SQLAlchemy. I can build a CREATE TABLE statement with a TEMPORARY clause by calling table._prefixes.append('TEMPORARY') against a Table object, but that's less elegant than table.select().prefix_with() used to add a prefix to data manipulation language expressions.
Is there an equivalent to .prefix_with() for DDL?
No, prefix_with() is defined for SELECT and INSERT only. But convenient way to add prefix to CREATE TABLE statement is passing it into table definition:
t = Table(
't', metadata,
Column('id', Integer, primary_key=True),
# ...
prefixes=['TEMPORARY'],
)
Related
The autoincrement argument in SQLAlchemy seems to be only True and False, but I want to set the pre-defined value aid = 1001, the via autoincrement aid = 1002 when the next insert is done.
In SQL, can be changed like:
ALTER TABLE article AUTO_INCREMENT = 1001;
I'm using MySQL and I have tried following, but it doesn't work:
from sqlalchemy.ext.declarative import declarative_base
Base = declarative_base()
class Article(Base):
__tablename__ = 'article'
aid = Column(INTEGER(unsigned=True, zerofill=True),
autoincrement=1001, primary_key=True)
So, how can I get that? Thanks in advance!
You can achieve this by using DDLEvents. This will allow you to run additional SQL statements just after the CREATE TABLE ran. Look at the examples in the link, but I am guessing your code will look similar to below:
from sqlalchemy import event
from sqlalchemy import DDL
event.listen(
Article.__table__,
"after_create",
DDL("ALTER TABLE %(table)s AUTO_INCREMENT = 1001;")
)
According to the docs:
autoincrement –
This flag may be set to False to indicate an integer primary key column that should not be considered to be the “autoincrement” column, that is the integer primary key column which generates values implicitly upon INSERT and whose value is usually returned via the DBAPI cursor.lastrowid attribute. It defaults to True to satisfy the common use case of a table with a single integer primary key column.
So, autoincrement is only a flag to let SQLAlchemy know whether it's the primary key you want to increment.
What you're trying to do is to create a custom autoincrement sequence.
So, your example, I think, should look something like:
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.schema import Sequence
Base = declarative_base()
class Article(Base):
__tablename__ = 'article'
aid = Column(INTEGER(unsigned=True, zerofill=True),
Sequence('article_aid_seq', start=1001, increment=1),
primary_key=True)
Note, I don't know whether you're using PostgreSQL or not, so you should make note of the following if you are:
The Sequence object also implements special functionality to accommodate Postgresql’s SERIAL datatype. The SERIAL type in PG automatically generates a sequence that is used implicitly during inserts. This means that if a Table object defines a Sequence on its primary key column so that it works with Oracle and Firebird, the Sequence would get in the way of the “implicit” sequence that PG would normally use. For this use case, add the flag optional=True to the Sequence object - this indicates that the Sequence should only be used if the database provides no other option for generating primary key identifiers.
I couldn't get the other answers to work using mysql and flask-migrate so I did the following inside a migration file.
from app import db
db.engine.execute("ALTER TABLE myDB.myTable AUTO_INCREMENT = 2000;")
Be warned that if you regenerated your migration files this will get overwritten.
I know this is an old question but I recently had to figure this out and none of the available answer were quite what I needed. The solution I found relied on Sequence in SQLAlchemy. For whatever reason, I could not get it to work when I called the Sequence constructor within the Column constructor as has been referenced above. As a note, I am using PostgreSQL.
For your answer I would have put it as such:
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy import Sequence, Column, Integer
import os
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy import Column, Sequence, Integer, create_engine
Base = declarative_base()
def connection():
engine = create_engine(f"postgresql://postgres:{os.getenv('PGPASSWORD')}#localhost:{os.getenv('PGPORT')}/test")
return engine
engine = connection()
class Article(Base):
__tablename__ = 'article'
seq = Sequence('article_aid_seq', start=1001)
aid = Column('aid', Integer, seq, server_default=seq.next_value(), primary_key=True)
Base.metadata.create_all(engine)
This then can be called in PostgreSQL with:
insert into article (aid) values (DEFAULT);
select * from article;
aid
------
1001
(1 row)
Hope this helps someone as it took me a while
You can do it using the mysql_auto_increment table create option. There are mysql_engine and mysql_default_charset options too, which might be also handy:
article = Table(
'article', metadata,
Column('aid', INTEGER(unsigned=True, zerofill=True), primary_key=True),
mysql_engine='InnoDB',
mysql_default_charset='utf8',
mysql_auto_increment='1001',
)
The above will generate:
CREATE TABLE article (
aid INTEGER UNSIGNED ZEROFILL NOT NULL AUTO_INCREMENT,
PRIMARY KEY (aid)
)ENGINE=InnoDB AUTO_INCREMENT=1001 DEFAULT CHARSET=utf8
If your database supports Identity columns*, the starting value can be set like this:
import sqlalchemy as sa
tbl = sa.Table(
't10494033',
sa.MetaData(),
sa.Column('id', sa.Integer, sa.Identity(start=200, always=True), primary_key=True),
)
Resulting in this DDL output:
CREATE TABLE t10494033 (
id INTEGER GENERATED ALWAYS AS IDENTITY (START WITH 200),
PRIMARY KEY (id)
)
Identity(..) is ignored if the backend does not support it.
* PostgreSQL 10+, Oracle 12+ and MSSQL, according to the linked documentation above.
I want create column ID via SQLAlchemy.
SQL seems like:
CREATE TABLE OBJECT (
ID INTEGER GENERATED BY DEFAULT AS IDENTITY PRIMARY KEY,
NAME VARCHAR(15)
)
Or how to create table with autoincrement field?
Column(autoincrement='auto')
default for primary_key, but that not work for Firebird.
I add fdb to github.
In this string SQL-code what needed for that, but i don't know how setup Column in SQLAlchemy to use that.
Screenshot from IBExpert showing what create after use "GENERATED BY DEFAULT AS IDENTITY". That not same as a simple generator.
This postgresql work around might work:
https://docs.sqlalchemy.org/en/13/dialects/postgresql.html
I believe you would have to modify the example for firebird by changing the compile dialect like this:
from sqlalchemy import MetaData
from sqlalchemy.schema import CreateColumn
from sqlalchemy.ext.compiler import compiles
#compiles(CreateColumn, 'firebird')
def use_identity(element, compiler, **kw):
text = compiler.visit_create_column(element, **kw)
text = text.replace("SERIAL", "INT GENERATED BY DEFAULT AS IDENTITY")
return text
m = MetaData()
t = Table(
't', m,
Column('id', Integer, primary_key=True),
Column('data', String)
)
I'm having difficulties writing what should be a simple SQL update statement in SQLAlchemy Core. However, I can't find any documentation, examples or tutorials that show how to combine multiple where conditions. I'm sure it's there - just can't find it.
Here's the table:
self.struct = Table('struct',
metadata,
Column('schema_name', String(40), nullable=False,
primary_key=True),
Column('struct_name', String(40), nullable=False,
primary_key=True),
Column('field_type', String(10), nullable=True),
Column('field_len', Integer, nullable=True) )
Here's the insert & update statement:
def struct_put(self, **kv):
try:
i = self.struct.insert()
result = i.execute(**kv)
except exc.IntegrityError: # row already exists - update it:
u = self.struct.update().\
where((self.struct.c.struct_name==kv['struct_name']
and self.struct.c.schema_name==kv['schema_name'])).\
values(field_len=kv['field_len'],
field_type=kv['field_type'])
result = u.execute()
The code handles the insert fine, but updates all rows in the table. Can you help me understand the syntax of this where clause? All suggestions are welcome - thanks in advance.
EDIT: The corrected clause looks like this:
where((and_(self.struct.c.parent_struct_name==kv['parent_struct_name'],
self.struct.c.struct_name==kv['struct_name'],
self.struct.c.schema_name==kv['schema_name']))).\
It's a very simple syntax, but given the many layers of SQLAlchemy it was surprisingly difficult to determine what exactly applied within this context.
It looks to me like you are using the Python "and" operation, which will evaluate to a only one of the clauses surrounding it. You should try using the "and_" function from SQLAlchemy instead. Put those two clauses inside the "and_" function.
You can also use & python operator.
For example:
query.where(
(ModelName.c.column_name == "column_value") &
(ModelName.c.column_name == "column_value)
)
For example, if you had a query like this
user_query = User.select().where(
(User.c.id == 12) &
(User.c.email == "myemail#gmail.com")
)
This will generate a raw SQL like this
select * from users where id = 12 and email = "myemail#gmail.com"
In SQLAlchemy, tablename.c is a special value that you use when constructing conditions that will be treated by SQLAlchemy at runtime.
In this particular case, you're simply saying "update all the rows where the column named struct_name matches the value passed in to struct_put(struct_name="struct_value", schema_name="schema_value"), and the column named schema_name matches the value passed in as schema_name.
Using table creation as normal:
t = Table(name, meta, [columns ...])
This is the first run where I create the table. In future executions I would like to use the table without having to indicate the [columns]. This seems redundant as it should already be specified in the table schema. In other words, for future accesses, I'd like to simply do:
t = Table(name, meta) # columns already read from schema
Is there a way to do this in SqlAlchemy?
See Reflecting Database Objects of SA documentation:
t = Table(name, meta, autoload=True)#, autoload_with=engine)
I'm using SQLAlchemy Migrate to keep track of database changes and I'm running into an issue with removing a foreign key. I have two tables, t_new is a new table, and t_exists is an existing table. I need to add t_new, then add a foreign key to t_exists. Then I need to be able to reverse the operation (which is where I'm having trouble).
t_new = sa.Table("new", meta.metadata,
sa.Column("new_id", sa.types.Integer, primary_key=True)
)
t_exists = sa.Table("exists", meta.metadata,
sa.Column("exists_id", sa.types.Integer, primary_key=True),
sa.Column(
"new_id",
sa.types.Integer,
sa.ForeignKey("new.new_id", onupdate="CASCADE", ondelete="CASCADE"),
nullable=False
)
)
This works fine:
t_new.create()
t_exists.c.new_id.create()
But this does not:
t_exists.c.new_id.drop()
t_new.drop()
Trying to drop the foreign key column gives an error: 1025, "Error on rename of '.\my_db_name\#sql-1b0_2e6' to '.\my_db_name\exists' (errno: 150)"
If I do this with raw SQL, i can remove the foreign key manually then remove the column, but I haven't been able to figure out how to remove the foreign key with SQLAlchemy? How can I remove the foreign key, and then the column?
You can do it with sqlalchemy.migrate.
In order to make it work, I have had to create the foreign key constraint explicitly rather than implicitely with Column('fk', ForeignKey('fk_table.field')):
Alas, instead of doing this:
p2 = Table('tablename',
metadata,
Column('id', Integer, primary_key=True),
Column('fk', ForeignKey('fk_table.field')),
mysql_engine='InnoDB',
)
do that:
p2 = Table('tablename',
metadata,
Column('id', Integer, primary_key=True),
Column('fk', Integer, index=True),
mysql_engine='InnoDB',
)
ForeignKeyConstraint(columns=[p2.c.fk], refcolumns=[p3.c.id]).create()
Then the deletion process looks like this:
def downgrade(migrate_engine):
# First drop the constraint
ForeignKeyConstraint(columns=[p2.c.fk], refcolumns=[p3.c.id]).drop()
# Then drop the table
p2.drop()
I was able to accomplish this by creating a separate metadata instance and using Session.execute() to run raw SQL. Ideally, there would be a solution that uses sqlalchemy exclusively, so I wouldn't have to use MySQL-specific solutions. But as of now, I am not aware of such a solution.
I believe you can achieve this with SQLAlchemy-Migrate. Note that a ForeignKey is on an isolated column. A ForeignKeyConstraint is at the table level and relates the columns together. If you look at the ForeignKey object on the column you will see that it references a ForeignKeyConstraint.
I would unable to test this idea because of the two databases I use MS SQL isn't supported by SqlAlchemy-Migrate and sqlite doesn't support "alter table" for constraints. I did get SQLAlchemy to try to remove a FK via a drop on the references constraint on a sqlite table so it was looking good. YMMV.
Well, you can achieve this in sqlalchemy: just drop() the all the constraints before you drop() the column (theoretically, you might have multiple constraints):
def drop_column(column):
for fk in column.table.foreign_keys:
if fk.column == column:
print 'deleting fk ', fk
fk.drop()
column.drop()
drop_column(t_exists.c.new_id)