The cursor class

class cursor

Allows Python code to execute PostgreSQL command in a database session. Cursors are created by the connection.cursor() method: they are bound to the connection for the entire lifetime and all the commands are executed in the context of the database session wrapped by the connection.

Cursors created from the same connection are not isolated, i.e., any changes done to the database by a cursor are immediately visible by the other cursors. Cursors created from different connections can or can not be isolated, depending on the connections’ isolation level. See also rollback() and commit() methods.

Cursors are not thread safe: a multithread application can create many cursors from the same connection and should use each cursor from a single thread. See Thread and process safety for details.

description

This read-only attribute is a sequence of 7-item sequences.

Each of these sequences is a named tuple (a regular tuple if collections.namedtuple() is not available) containing information describing one result column:

  1. name: the name of the column returned.
  2. type_code: the PostgreSQL OID of the column. You can use the pg_type system table to get more informations about the type. This is the value used by Psycopg to decide what Python type use to represent the value. See also Type casting of SQL types into Python objects.
  3. display_size: the actual length of the column in bytes. Obtaining this value is computationally intensive, so it is always None unless the PSYCOPG_DISPLAY_SIZE parameter is set at compile time. See also PQgetlength.
  4. internal_size: the size in bytes of the column associated to this column on the server. Set to a negative value for variable-size types See also PQfsize.
  5. precision: total number of significant digits in columns of type NUMERIC. None for other types.
  6. scale: count of decimal digits in the fractional part in columns of type NUMERIC. None for other types.
  7. null_ok: always None as not easy to retrieve from the libpq.

This attribute will be None for operations that do not return rows or if the cursor has not had an operation invoked via the execute*() methods yet.

Changed in version 2.4: if possible, columns descriptions are named tuple instead of regular tuples.

close()

Close the cursor now (rather than whenever del is executed). The cursor will be unusable from this point forward; an InterfaceError will be raised if any operation is attempted with the cursor.

Changed in version 2.5: if the cursor is used in a with statement, the method is automatically called at the end of the with block.

closed

Read-only boolean attribute: specifies if the cursor is closed (True) or not (False).

DB API extension

The closed attribute is a Psycopg extension to the DB API 2.0.

New in version 2.0.7.

connection

Read-only attribute returning a reference to the connection object on which the cursor was created.

name

Read-only attribute containing the name of the cursor if it was creates as named cursor by connection.cursor(), or None if it is a client side cursor. See Server side cursors.

DB API extension

The name attribute is a Psycopg extension to the DB API 2.0.

scrollable

Read/write attribute: specifies if a named cursor is declared SCROLL, hence is capable to scroll backwards (using scroll()). If True, the cursor can be scrolled backwards, if False it is never scrollable. If None (default) the cursor scroll option is not specified, usually but not always meaning no backward scroll (see the DECLARE notes).

Note

set the value before calling execute() or use the connection.cursor() scrollable parameter, otherwise the value will have no effect.

New in version 2.5.

DB API extension

The scrollable attribute is a Psycopg extension to the DB API 2.0.

withhold

Read/write attribute: specifies if a named cursor lifetime should extend outside of the current transaction, i.e., it is possible to fetch from the cursor even after a connection.commit() (but not after a connection.rollback()). See Server side cursors

Note

set the value before calling execute() or use the connection.cursor() withhold parameter, otherwise the value will have no effect.

New in version 2.4.3.

DB API extension

The withhold attribute is a Psycopg extension to the DB API 2.0.

Commands execution methods

execute(operation[, parameters])

Prepare and execute a database operation (query or command).

Parameters may be provided as sequence or mapping and will be bound to variables in the operation. Variables are specified either with positional (%s) or named (%(name)s) placeholders. See Passing parameters to SQL queries.

The method returns None. If a query was executed, the returned values can be retrieved using fetch*() methods.

executemany(operation, seq_of_parameters)

Prepare a database operation (query or command) and then execute it against all parameter tuples or mappings found in the sequence seq_of_parameters.

The function is mostly useful for commands that update the database: any result set returned by the query is discarded.

Parameters are bounded to the query using the same rules described in the execute() method.

callproc(procname[, parameters])

Call a stored database procedure with the given name. The sequence of parameters must contain one entry for each argument that the procedure expects. The result of the call is returned as modified copy of the input sequence. Input parameters are left untouched, output and input/output parameters replaced with possibly new values.

The procedure may also provide a result set as output. This must then be made available through the standard fetch*() methods.

mogrify(operation[, parameters])

Return a query string after arguments binding. The string returned is exactly the one that would be sent to the database running the execute() method or similar.

The returned string is always a bytes string.

>>> cur.mogrify("INSERT INTO test (num, data) VALUES (%s, %s)", (42, 'bar'))
"INSERT INTO test (num, data) VALUES (42, E'bar')"

DB API extension

The mogrify() method is a Psycopg extension to the DB API 2.0.

setinputsizes(sizes)

This method is exposed in compliance with the DB API 2.0. It currently does nothing but it is safe to call it.

Results retrieval methods

The following methods are used to read data from the database after an execute() call.

Note

cursor objects are iterable, so, instead of calling explicitly fetchone() in a loop, the object itself can be used:

>>> cur.execute("SELECT * FROM test;")
>>> for record in cur:
...     print record
...
(1, 100, "abc'def")
(2, None, 'dada')
(3, 42, 'bar')

Changed in version 2.4: iterating over a named cursor fetches itersize records at time from the backend. Previously only one record was fetched per roundtrip, resulting in a large overhead.

fetchone()

Fetch the next row of a query result set, returning a single tuple, or None when no more data is available:

>>> cur.execute("SELECT * FROM test WHERE id = %s", (3,))
>>> cur.fetchone()
(3, 42, 'bar')

A ProgrammingError is raised if the previous call to execute*() did not produce any result set or no call was issued yet.

fetchmany([size=cursor.arraysize])

Fetch the next set of rows of a query result, returning a list of tuples. An empty list is returned when no more rows are available.

The number of rows to fetch per call is specified by the parameter. If it is not given, the cursor’s arraysize determines the number of rows to be fetched. The method should try to fetch as many rows as indicated by the size parameter. If this is not possible due to the specified number of rows not being available, fewer rows may be returned:

>>> cur.execute("SELECT * FROM test;")
>>> cur.fetchmany(2)
[(1, 100, "abc'def"), (2, None, 'dada')]
>>> cur.fetchmany(2)
[(3, 42, 'bar')]
>>> cur.fetchmany(2)
[]

A ProgrammingError is raised if the previous call to execute*() did not produce any result set or no call was issued yet.

Note there are performance considerations involved with the size parameter. For optimal performance, it is usually best to use the arraysize attribute. If the size parameter is used, then it is best for it to retain the same value from one fetchmany() call to the next.

fetchall()

Fetch all (remaining) rows of a query result, returning them as a list of tuples. An empty list is returned if there is no more record to fetch.

>>> cur.execute("SELECT * FROM test;")
>>> cur.fetchall()
[(1, 100, "abc'def"), (2, None, 'dada'), (3, 42, 'bar')]

A ProgrammingError is raised if the previous call to execute*() did not produce any result set or no call was issued yet.

scroll(value[, mode='relative'])

Scroll the cursor in the result set to a new position according to mode.

If mode is relative (default), value is taken as offset to the current position in the result set, if set to absolute, value states an absolute target position.

If the scroll operation would leave the result set, a ProgrammingError is raised and the cursor position is not changed.

Note

According to the DB API 2.0, the exception raised for a cursor out of bound should have been IndexError. The best option is probably to catch both exceptions in your code:

try:
    cur.scroll(1000 * 1000)
except (ProgrammingError, IndexError), exc:
    deal_with_it(exc)

The method can be used both for client-side cursors and server-side cursors. Server-side cursors can usually scroll backwards only if declared scrollable. Moving out-of-bound in a server-side cursor doesn’t result in an exception, if the backend doesn’t raise any (Postgres doesn’t tell us in a reliable way if we went out of bound).

arraysize

This read/write attribute specifies the number of rows to fetch at a time with fetchmany(). It defaults to 1 meaning to fetch a single row at a time.

itersize

Read/write attribute specifying the number of rows to fetch from the backend at each network roundtrip during iteration on a named cursor. The default is 2000.

New in version 2.4.

DB API extension

The itersize attribute is a Psycopg extension to the DB API 2.0.

rowcount

This read-only attribute specifies the number of rows that the last execute*() produced (for DQL statements like SELECT) or affected (for DML statements like UPDATE or INSERT).

The attribute is -1 in case no execute*() has been performed on the cursor or the row count of the last operation if it can’t be determined by the interface.

Note

The DB API 2.0 interface reserves to redefine the latter case to have the object return None instead of -1 in future versions of the specification.

rownumber

This read-only attribute provides the current 0-based index of the cursor in the result set or None if the index cannot be determined.

The index can be seen as index of the cursor in a sequence (the result set). The next fetch operation will fetch the row indexed by rownumber in that sequence.

lastrowid

This read-only attribute provides the OID of the last row inserted by the cursor. If the table wasn’t created with OID support or the last operation is not a single record insert, the attribute is set to None.

Note

PostgreSQL currently advices to not create OIDs on the tables and the default for CREATE TABLE is to not support them. The INSERT ... RETURNING syntax available from PostgreSQL 8.3 allows more flexibility.

query

Read-only attribute containing the body of the last query sent to the backend (including bound arguments) as bytes string. None if no query has been executed yet:

>>> cur.execute("INSERT INTO test (num, data) VALUES (%s, %s)", (42, 'bar'))
>>> cur.query
"INSERT INTO test (num, data) VALUES (42, E'bar')"

DB API extension

The query attribute is a Psycopg extension to the DB API 2.0.

statusmessage

Read-only attribute containing the message returned by the last command:

>>> cur.execute("INSERT INTO test (num, data) VALUES (%s, %s)", (42, 'bar'))
>>> cur.statusmessage
'INSERT 0 1'

DB API extension

The statusmessage attribute is a Psycopg extension to the DB API 2.0.

cast(oid, s)

Convert a value from the PostgreSQL string representation to a Python object.

Use the most specific of the typecasters registered by register_type().

New in version 2.4.

DB API extension

The cast() method is a Psycopg extension to the DB API 2.0.

tzinfo_factory

The time zone factory used to handle data types such as TIMESTAMP WITH TIME ZONE. It should be a tzinfo object. A few implementations are available in the psycopg2.tz module.

nextset()

This method is not supported (PostgreSQL does not have multiple data sets) and will raise a NotSupportedError exception.

setoutputsize(size[, column])

This method is exposed in compliance with the DB API 2.0. It currently does nothing but it is safe to call it.

COPY-related methods

Efficiently copy data from file-like objects to the database and back. See Using COPY TO and COPY FROM for an overview.

DB API extension

The COPY command is a PostgreSQL extension to the SQL standard. As such, its support is a Psycopg extension to the DB API 2.0.

copy_from(file, table, sep='\t', null='\\N', size=8192, columns=None)

Read data from the file-like object file appending them to the table named table.

Parameters:
  • file – file-like object to read data from. It must have both read() and readline() methods.
  • table – name of the table to copy data into.
  • sep – columns separator expected in the file. Defaults to a tab.
  • null – textual representation of NULL in the file. The default is the two characters string \N.
  • size – size of the buffer used to read from the file.
  • columns – iterable with name of the columns to import. The length and types should match the content of the file to read. If not specified, it is assumed that the entire table matches the file structure.

Example:

>>> f = StringIO("42\tfoo\n74\tbar\n")
>>> cur.copy_from(f, 'test', columns=('num', 'data'))
>>> cur.execute("select * from test where id > 5;")
>>> cur.fetchall()
[(6, 42, 'foo'), (7, 74, 'bar')]

Note

the name of the table is not quoted: if the table name contains uppercase letters or special characters it must be quoted with double quotes:

cur.copy_from(f, '"TABLE"')

Changed in version 2.0.6: added the columns parameter.

Changed in version 2.4: data read from files implementing the io.TextIOBase interface are encoded in the connection encoding when sent to the backend.

copy_to(file, table, sep='\t', null='\\N', columns=None)

Write the content of the table named table to the file-like object file. See Using COPY TO and COPY FROM for an overview.

Parameters:
  • file – file-like object to write data into. It must have a write() method.
  • table – name of the table to copy data from.
  • sep – columns separator expected in the file. Defaults to a tab.
  • null – textual representation of NULL in the file. The default is the two characters string \N.
  • columns – iterable with name of the columns to export. If not specified, export all the columns.

Example:

>>> cur.copy_to(sys.stdout, 'test', sep="|")
1|100|abc'def
2|\N|dada
...

Note

the name of the table is not quoted: if the table name contains uppercase letters or special characters it must be quoted with double quotes:

cur.copy_to(f, '"TABLE"')

Changed in version 2.0.6: added the columns parameter.

Changed in version 2.4: data sent to files implementing the io.TextIOBase interface are decoded in the connection encoding when read from the backend.

copy_expert(sql, file, size=8192)

Submit a user-composed COPY statement. The method is useful to handle all the parameters that PostgreSQL makes available (see COPY command documentation).

Parameters:
  • sql – the COPY statement to execute.
  • file – a file-like object to read or write (according to sql).
  • size – size of the read buffer to be used in COPY FROM.

The sql statement should be in the form COPY table TO STDOUT to export table to the file object passed as argument or COPY table FROM STDIN to import the content of the file object into table.

file must be a readable file-like object (as required by copy_from()) for sql statement COPY ... FROM STDIN or a writable one (as required by copy_to()) for COPY ... TO STDOUT.

Example:

>>> cur.copy_expert("COPY test TO STDOUT WITH CSV HEADER", sys.stdout)
id,num,data
1,100,abc'def
2,,dada
...

New in version 2.0.6.

Changed in version 2.4: files implementing the io.TextIOBase interface are dealt with using Unicode data instead of bytes.