FDB - The Python driver for Firebird¶

Installation from PYPI¶

Run pip:

$ pip install fdb

Installation from source¶

Download the source tarball, uncompress it, then run the install command:

$ tar -xzvf fdb-2.0.tar.gz
$ cd fdb-2.0
$ python setup.py install

Connecting to a Database¶

Example 1

A database connection is typically established with code such as this:

import fdb

# The server is named 'bison'; the database file is at '/temp/test.db'.
con = fdb.connect(dsn='bison:/temp/test.db', user='sysdba', password='pass')

# Or, equivalently:
con = fdb.connect(
    host='bison', database='/temp/test.db',
    user='sysdba', password='pass'
  )

Example 2

Suppose we want to connect to the database in SQL Dialect 1 and specifying UTF-8 as the character set of the connection:

import fdb

con = fdb.connect(
    dsn='bison:/temp/test.db',
    user='sysdba', password='pass',
    dialect=1, # necessary for all dialect 1 databases
    charset='UTF8' # specify a character set for the connection
  )

Executing SQL Statements¶

For this section, suppose we have a table defined and populated by the following SQL code:

create table languages
(
  name               varchar(20),
  year_released      integer
);

insert into languages (name, year_released) values ('C',        1972);
insert into languages (name, year_released) values ('Python',   1991);

Example 1

This example shows the simplest way to print the entire contents of the languages table:

import fdb

con = fdb.connect(dsn='/temp/test.db', user='sysdba', password='masterkey')

# Create a Cursor object that operates in the context of Connection con:
cur = con.cursor()

# Execute the SELECT statement:
cur.execute("select * from languages order by year_released")

# Retrieve all rows as a sequence and print that sequence:
print cur.fetchall()

Sample output:

[('C', 1972), ('Python', 1991)]

Example 2

Here’s another trivial example that demonstrates various ways of fetching a single row at a time from a SELECT-cursor:

import fdb

con = fdb.connect(dsn='/temp/test.db', user='sysdba', password='masterkey')

cur = con.cursor()
SELECT = "select name, year_released from languages order by year_released"

# 1. Iterate over the rows available from the cursor, unpacking the
# resulting sequences to yield their elements (name, year_released):
cur.execute(SELECT)
for (name, year_released) in cur:
    print '%s has been publicly available since %d.' % (name, year_released)

# 2. Equivalently:
cur.execute(SELECT)
for row in cur:
    print '%s has been publicly available since %d.' % (row[0], row[1])

# 3. Using mapping-iteration rather than sequence-iteration:
cur.execute(SELECT)
for row in cur.itermap():
    print '%(name)s has been publicly available since %(year_released)d.' % row

Sample output:

C has been publicly available since 1972.
Python has been publicly available since 1991.
C has been publicly available since 1972.
Python has been publicly available since 1991.
C has been publicly available since 1972.
Python has been publicly available since 1991.

Example 3

The following program is a simplistic table printer (applied in this example to languages):

import fdb

TABLE_NAME = 'languages'
SELECT = 'select * from %s order by year_released' % TABLE_NAME

con = fdb.connect(dsn='/temp/test.db', user='sysdba', password='masterkey')

cur = con.cursor()
cur.execute(SELECT)

# Print a header.
for fieldDesc in cur.description:
    print fieldDesc[fdb.DESCRIPTION_NAME].ljust(fieldDesc[fdb.DESCRIPTION_DISPLAY_SIZE]) ,
print # Finish the header with a newline.
print '-' * 78

# For each row, print the value of each field left-justified within
# the maximum possible width of that field.
fieldIndices = range(len(cur.description))
for row in cur:
    for fieldIndex in fieldIndices:
        fieldValue = str(row[fieldIndex])
        fieldMaxWidth = cur.description[fieldIndex][fdb.DESCRIPTION_DISPLAY_SIZE]

        print fieldValue.ljust(fieldMaxWidth) ,

    print # Finish the row with a newline.

Sample output:

NAME                 YEAR_RELEASED
------------------------------------------------------------------------------
C                    1972
Python               1991

Example 4

Let’s insert more languages:

import fdb

con = fdb.connect(dsn='/temp/test.db', user='sysdba', password='masterkey')

cur = con.cursor()

newLanguages = [
    ('Lisp',  1958),
    ('Dylan', 1995),
  ]

cur.executemany("insert into languages (name, year_released) values (?, ?)",
    newLanguages
  )

# The changes will not be saved unless the transaction is committed explicitly:
con.commit()

Note the use of a parameterized SQL statement above. When dealing with repetitive statements, this is much faster and less error-prone than assembling each SQL statement manually. (You can read more about parameterized SQL statements in the section on Prepared Statements.)

After running Example 4, the table printer from Example 3 would print:

NAME                 YEAR_RELEASED
------------------------------------------------------------------------------
Lisp                 1958
C                    1972
Python               1991
Dylan                1995

Calling Stored Procedures¶

Firebird supports stored procedures written in a proprietary procedural SQL language. Firebird stored procedures can have input parameters and/or output parameters. Some databases support input/output parameters, where the same parameter is used for both input and output; Firebird does not support this.

It is important to distinguish between procedures that return a result set and procedures that populate and return their output parameters exactly once. Conceptually, the latter “return their output parameters” like a Python function, whereas the former “yield result rows” like a Python generator.

Firebird’s server-side procedural SQL syntax makes no such distinction, but client-side SQL code (and C API code) must. A result set is retrieved from a stored procedure by SELECT`ing from the procedure, whereas output parameters are retrieved with an `EXECUTE PROCEDURE statement.

To retrieve a result set from a stored procedure with FDB, use code such as this:

cur.execute("select output1, output2 from the_proc(?, ?)", (input1, input2))

# Ordinary fetch code here, such as:
for row in cur:
    ... # process row

con.commit() # If the procedure had any side effects, commit them.

To execute a stored procedure and access its output parameters, use code such as this:

cur.callproc("the_proc", (input1, input2))

# If there are output parameters, retrieve them as though they were the
# first row of a result set.  For example:
outputParams = cur.fetchone()

con.commit() # If the procedure had any side effects, commit them.

This latter is not very elegant; it would be preferable to access the procedure’s output parameters as the return value of Cursor.callproc(). The Python DB API specification requires the current behavior, however.

Using connect¶

This constructor has number of keyword parameters that could be divided into several groups:

• Database specification (parameters dsn, host, database and port)
• User specification (parameters user, password and role)
• Connection options (parameters sql_dialect, charset, isolation_level, buffers, force_writes, no_reserve and db_key_scope)

To establish a connection to database, you always must specify the database, either as connection string parameter dsn, or as required combination of parameters host, database and port.

Although specification of user and password parameters is optional (if environment variables ISC_USER and ISC_PASSWORD are set, their values are used if these parameters are ommited), it’s recommended practice to use them. Parameter role is needed only when you use Firebird roles.

Connection options are optional (see Firebird Documentation for details). However you may often want to specify charset, as it directs automatic conversions of string data between client and server, and automatic conversions from/to unicode performed by FDB driver (see Data handling and conversions for details).

Examples:

# Connecting via 'dsn'
#
# Local database (local protocol, if supported)
con = fdb.connect(dsn='/path/database.fdb', user='sysdba', password='pass')
# Local database (TCP/IP)
con = fdb.connect(dsn='localhost:/path/database.fdb', user='sysdba', password='pass')
# Local database (TCP/IP with port specification)
con = fdb.connect(dsn='localhost/3050:/path/database.fdb', user='sysdba', password='pass')
# Remote database
con = fdb.connect(dsn='host:/path/database.db', user='sysdba', password='pass')
# Remote database with port specification
con = fdb.connect(dsn='host/3050:/path/database.db', user='sysdba', password='pass')
#
# Connecting via 'database', 'host' and 'port'
#
# Local database (local protocol, if supported)
con = fdb.connect(database='/path/database.db', user='sysdba', password='pass')
# Local database (TCP/IP)
con = fdb.connect(host='localhost', database='/path/database.db', user='sysdba', password='pass')
# Local database (TCP/IP with port specification)
con = fdb.connect(host='localhost', port=3050, database='/path/database.db', user='sysdba', password='pass')
# Remote database
con = fdb.connect(host='myhost', database='/path/database.db', user='sysdba', password='pass')

Since version 1.2 FDB supports additional Connection class(es) that extend Connection functionality in optional (opt-in) way. For example ConnectionWithSchema extends Connection interface with methods and attributes provided by Schema. New connection_class parameter was introduced to connect and create_database to connect to/create database using different class than descends from Connection.

Example:

# Connecting through ConnectionWithSchema
#
con = fdb.connect(dsn='/path/database.fdb', user='sysdba', password='pass',
                  connection_class=fdb.ConnectionWithSchema)

Using create_database¶

The Firebird engine supports dynamic database creation via the SQL statement CREATE DATABASE. FDB wraps it into create_database(), that returns Connection instance attached to newly created database.

Example:

con = fdb.create_database("create database 'host:/temp/db.db' user 'sysdba' password 'pass'")

Example:

con = fdb.create_database(dsn='/temp/db.fdb',user='sysdba',password='pass',page_size=8192)

Deleting databases¶

The Firebird engine also supports dropping (deleting) databases dynamically, but dropping is a more complicated operation than creating, for several reasons: an existing database may be in use by users other than the one who requests the deletion, it may have supporting objects such as temporary sort files, and it may even have dependent shadow databases. Although the database engine recognizes a DROP DATABASE SQL statement, support for that statement is limited to the isql command-line administration utility. However, the engine supports the deletion of databases via an API call, which FDB exposes as drop_database() method in Connection class. So, to drop a database you need to connect to it first.

Examples:

import fdb

con = fdb.create_database("create database '/temp/db.db' user 'sysdba' password 'pass'")
con.drop_database()

con = fdb.connect(dsn='/path/database.fdb', user='sysdba', password='pass')
con.drop_database()

Connection object¶

Connection object represents a direct link to database, and works as gateway for next operations with it:

• Executing SQL Statements: methods execute_immediate() and cursor().
• Dropping database: method drop_database().
• Transanction management: methods begin(), commit(), rollback(), savepoint(), trans(), trans_info() and transaction_info(), and attributes main_transaction, transactions, default_tpb and group.
• Work with Database Events: method event_conduit().
• Getting information about Firebird version: attributes server_version, firebird_version, version and engine_version.
• Getting information about database: methods db_info() and database_info().
• Getting information about database metadata: attribute schema and ods.

Getting information about Firebird version¶

Because functionality and some features depends on actual Firebird version, it could be important for FDB users to check it. This (otherwise) simple task could be confusing for new Firebird users, because Firebird uses two different version lineages. This abomination was introduced to Firebird thanks to its InterBase legacy (Firebird 1.0 is a fork of InterBase 6.0), as applications designed to work with InterBase can often work with Firebird without problems (and vice versa). However, legacy applications designed to work with InterBase may stop working properly if they would detect unexpectedly low server version, so default version number returned by Firebird (and FDB) is based on InterBase version number. For example this version for Firebird 2.5.2 is 6.3.2, so condition for legacy applications that require at least IB 6.0 is met.

FDB provides these version strings as two Connection properties:

• server_version - Legacy InterBase-friendly version string.
• firebird_version - Firebird’s own version string.

However, this version string contains more information than version number. For example for Linux Firebird 2.5.2 it’s ‘LI-V2.5.2.26540 Firebird 2.5’. So FDB provides two more properties for your convenience:

• version - Only Firebird version number. It’s a string with format: major.minor.subrelease.build
• engine_version - Engine (major.minor) version as (float) number.

FDB also provides convenient constants for supported engine versions: ODS_FB_20,`ODS_FB_21` and ODS_FB_25.

Database On-Disk Structure¶

Particular Firebird features may also depend on specific support in database (for example monitoring tables introduced in Firebird 2.1). These required structures are present automatically when database is created by particular engine verison that needs them, but Firebird engine may typically work with databases created by older versions and thus with older structure, so it could be necessary to consult also On-Disk Structure (ODS for short) version. FDB provides this number as Connection.ods (float) property.

Example:

con = fdb.connect(dsn='/path/database.fdb', user='sysdba', password='pass')
print 'Firebird version:',con.version
print 'ODS version:',con.ods

Firebird version: 2.5.2.26540
ODS version: 11.1

In abover example although connected Firebird engine is version 2.5, connected database has ODS 11.1 which came with Firebird 2.1, and some Firebird 2.5 features will not be available on this database.

Getting information about database¶

Firebird provides various informations about server and connected database via database_info API call. FDB surfaces this API through methods db_info() and database_info() on Connection object.

Connection.database_info() is a very thin wrapper around function isc_database_info(). This method does not attempt to interpret its results except with regard to whether they are a string or an integer. For example, requesting isc_info_user_names with the call:

con.database_info(fdb.isc_info_user_names, 's')

will return a binary string containing a raw succession of length-name pairs.

Example program:

import fdb

con = fdb.connect(dsn='localhost:/temp/test.db', user='sysdba', password='pass')

# Retrieving an integer info item is quite simple.
bytesInUse = con.database_info(fdb.isc_info_current_memory, 'i')

print 'The server is currently using %d bytes of memory.' % bytesInUse

# Retrieving a string info item is somewhat more involved, because the
# information is returned in a raw binary buffer that must be parsed
# according to the rules defined in the Interbase® 6 API Guide section
# entitled "Requesting buffer items and result buffer values" (page 51).
#
# Often, the buffer contains a succession of length-string pairs
# (one byte telling the length of s, followed by s itself).
# Function fdb.ibase.ord2 is provided to convert a raw
# byte to a Python integer (see examples below).
buf = con.database_info(fdb.isc_info_db_id, 's')

# Parse the filename from the buffer.
beginningOfFilename = 2
# The second byte in the buffer contains the size of the database filename
# in bytes.
lengthOfFilename = fdb.ibase.ord2(buf[1])
filename = buf[beginningOfFilename:beginningOfFilename + lengthOfFilename]

# Parse the host name from the buffer.
beginningOfHostName = (beginningOfFilename + lengthOfFilename) + 1
# The first byte after the end of the database filename contains the size
# of the host name in bytes.
lengthOfHostName = fdb.ibase.ord2(buf[beginningOfHostName - 1])
host = buf[beginningOfHostName:beginningOfHostName + lengthOfHostName]

print 'We are connected to the database at %s on host %s.' % (filename, host)

Sample output:

The server is currently using 8931328 bytes of memory.
We are connected to the database at C:\TEMP\TEST.DB on host WEASEL.

A more convenient way to access the same functionality is via the db_info() method, which is high-level convenience wrapper around the database_info() method that parses the output of database_info into Python-friendly objects instead of returning raw binary buffers in the case of complex result types. For example, requesting isc_info_user_names with the call:

con.db_info(fdb.isc_info_user_names)

returns a dictionary that maps (username -> number of open connections). If SYSDBA has one open connection to the database to which con is connected, and TEST_USER_1 has three open connections to that same database, the return value would be:

{‘SYSDBA’: 1, ‘TEST_USER_1’: 3}

Example program:

import fdb
import os.path

###############################################################################
# Querying an isc_info_* item that has a complex result:
###############################################################################
# Establish three connections to the test database as TEST_USER_1, and one
# connection as SYSDBA.  Then use the Connection.db_info method to query the
# number of attachments by each user to the test database.
testUserCons = []
for i in range(3):
  tcon = fdb.connect(dsn='localhost:/temp/test.db', user='TEST_USER_1', password='pass')
  testUserCons.append(tcon)

con = fdb.connect(dsn='localhost:/temp/test.db', user='sysdba', password='pass')

print 'Open connections to this database:'
print con.db_info(fdb.isc_info_user_names)

###############################################################################
# Querying multiple isc_info_* items at once:
###############################################################################
# Request multiple db_info items at once, specifically the page size of the
# database and the number of pages currently allocated.  Compare the size
# computed by that method with the size reported by the file system.
# The advantages of using db_info instead of the file system to compute
# database size are:
#   - db_info works seamlessly on connections to remote databases that reside
#     in file systems to which the client program lacks access.
#   - If the database is split across multiple files, db_info includes all of
#     them.
res = con.db_info([fdb.isc_info_page_size, fdb.isc_info_allocation])
pagesAllocated = res[fdb.isc_info_allocation]
pageSize = res[fdb.isc_info_page_size]
print '\ndb_info indicates database size is', pageSize * pagesAllocated, 'bytes'
print 'os.path.getsize indicates size is ', os.path.getsize(DB_FILENAME), 'bytes'

Sample output:

Open connections to this database:
{'SYSDBA': 1, 'TEST_USER_1': 3}

db_info indicates database size is 20684800 bytes
os.path.getsize indicates size is  20684800 bytes

Cursor object¶

Because Cursor objects always operate in context of single Connection (and Transaction), Cursor instances are not created directly, but by constructor method. Python DB API 2.0 assume that if database engine supports transactions, it supports only one transaction per connection, hence it defines constructor method cursor() (and other transaction-related methods) as part of Connection interface. However, Firebird supports multiple independent transactions per connection. To conform to Python DB API, FDB uses concept of internal main_transaction and secondary transactions. Cursor constructor is primarily defined by Transaction, and Cursor constructor on Connection is therefore a shortcut for main_transaction.cursor().

Cursor objects are used for next operations:

• Execution of SQL Statemets: methods execute(), executemany() and callproc().
• Creating PreparedStatement objects for efficient repeated execution of SQL statements, and to obtain additional information about SQL statements (like execution plan): method prep().
• Fetching results: methods fetchone(), fetchmany(), fetchall(), fetchonemap(), fetchmanymap(), fetchallmap(), iter(), itermap() and next().

SQL Execution Basics¶

There are three methods how to execute SQL commands:

• Connection.execute_immediate() or Transaction.execute_immediate() for SQL commands that don’t return any result, and are not executed frequently. This method also doesn’t support either parametrized statements or prepared statements.

  Tip

  This method is efficient for administrative and DDL SQL commands, like DROP, CREATE or ALTER commands, SET STATISTICS etc.

• Cursor.execute() or Cursor.executemany() for commands that return result sets, i.e. sequence of rows of the same structure, and sequence has unknown number of rows (including zero).

  Tip

  This method is preferred for all SELECT and other DML statements, or any statement that is executed frequently, either as is or in parametrized form.

• Cursor.callproc() for execution of Stored procedures that always return exactly one set of values.

  Note

  This method of SP invocation is equivalent to “EXECUTE PROCEDURE …” SQL statement.

Parametrized statements¶

When SQL command you want to execute contains data values, you can either:

• Embed them directly or via string formatting into command string, e.g.:

  cur.execute("insert into the_table (a,b,c) values ('aardvark', 1, 0.1)")
  # or
  cur.execute("select * from the_table where col == 'aardvark'")
  # or
  cur.execute("insert into the_table (a,b,c) values ('%s', %i, %f)" % ('aardvark',1,0.1))
  # or
  cur.execute("select * from the_table where col == '%s'" % 'aardvark')
  

• Use parameter marker (?) in command string in the slots where values are expected, then supply those values as Python list or tuple:

  cur.execute("insert into the_table (a,b,c) values (?,?,?)", ('aardvark', 1, 0.1))
  # or
  cur.execute("select * from the_table where col == ?",('aardvark',))
  

While both methods have the same results, the second one (called parametrized) has several important advantages:

• You don’t need to handle conversions from Python data types to strings.
• FDB will handle all data type conversions (if necessary) from Python data types to Firebird ones, including None/NULL conversion and conversion from unicode to byte strings in encoding expected by server.
• You may pass BLOB values as open file-like objects, and FDB will handle the transfer of BLOB value.

Parametrized statemets also have some limitations. Currently:

• DATE, TIME and DATETIME values must be relevant datetime objects.
• NUMERIC and DECIMAL values must be decimal objects.

Fetching data from server¶

Result of SQL statement execution consists from sequence of zero to unknown number of rows, where each row is a set of exactly the same number of values. Cursor object offer number of different methods for fetching these rows, that should satisfy all your specific needs:

• fetchone() - Returns the next row of a query result set, or None when no more data is available.

  Tip

  Cursor supports the iterator protocol, yielding tuples of values like fetchone().

• fetchmany() - Returns the next set of rows of a query result, returning a sequence of sequences (e.g. a list of tuples). An empty sequence 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 does 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.

  Note

  The default value of arraysize is 1, so without paremeter it’s equivalent to fetchone(), but returns list of rows, instead actual row directly.

• fetchall() - Returns all (remaining) rows of a query result as list of tuples, where each tuple is one row of returned values.

  Tip

  This method can potentially return huge amount of data, that may exhaust available memory. If you need just iteration over potentially big result set, use loops with fetchone(), Cursor’s built-in support for iterator protocol or call to iter() instead this method.

• fetchonemap() - Returns the next row like fetchone(), but returns a mapping of field name to field value, rather than a tuple.

• fetchmanymap() - Returns the next set of rows of a query result like fetchmany(), but returns a list of mapping of field name to field value, rather than a tuple.

• fetchallmap() - Returns all (remaining) rows of a query result like fetchall(), returns a list of mappings of field name to field value, rather than a tuple.

  Tip

  This method can potentially return huge amount of data, that may exhaust available memory. If you need just iteration over potentially big result set with mapping support, use itermap() instead this method.

• iter() - Equivalent to the fetchall(), except that it returns iterator rather than materialized list.

• itermap() - Equivalent to the fetchallmap(), except that it returns iterator rather than materialized list.

• Call to execute() returns self (Cursor instance) that itself supports the iterator protocol, yielding tuples of values like fetchone().

Examples:

import fdb

con = fdb.connect(dsn='/temp/test.db', user='sysdba', password='masterkey')

cur = con.cursor()
SELECT = "select name, year_released from languages order by year_released"

# 1. Using built-in support for iteration protocol to iterate over the rows available from the cursor,
# unpacking the resulting sequences to yield their elements (name, year_released):
cur.execute(SELECT)
for (name, year_released) in cur:
    print '%s has been publicly available since %d.' % (name, year_released)
# or alternatively you can take an advantage of cur.execute returning self.
for (name, year_released) in cur.execute(SELECT):
    print '%s has been publicly available since %d.' % (name, year_released)

# 2. Equivalently using fetchall():
# This is potentially dangerous if result set is huge, as the whole result set is first materialized
# as list and then used for iteration.
cur.execute(SELECT)
for row in cur.fetchall():
    print '%s has been publicly available since %d.' % (row[0], row[1])

# 3. Using mapping-iteration rather than sequence-iteration:
cur.execute(SELECT)
for row in cur.itermap():
    print '%(name)s has been publicly available since %(year_released)d.' % row

Prepared Statements¶

Execution of any SQL statement has three phases:

• Preparation: command is analyzed, validated, execution plan is determined by optimizer and all necessary data structures (for example for input and output parameters) are initialized.
• Execution: input parameters (if any) are passed to server and previously prepared statement is actually executed by database engine.
• Fetching: result of execution and data (if any) are transfered from server to client, and allocated resources are then released.

The preparation phase consumes some amount of server resources (memory and CPU). Although preparation and release of resources typically takes only small amount of CPU time, it builds up as number of executed statements grows. Firebird (like most database engines) allows to spare this time for subsequent execution if particular statement should be executed repeatedly - by reusing once prepared statement for repeated execution. This may save significant amount of server processing time, and result in better overall performance.

FDB builds on this by encapsulating all statement-related code into separate PreparedStatement class, and implementing Cursor class as a wrapper around it.

Beside SQL command string, Cursor also allows to aquire and use PreparedStatement instances explicitly. PreparedStatement are aquired by calling prep() method could be then passed to execute() or executemany() instead command string.

Example:

insertStatement = cur.prep("insert into the_table (a,b,c) values (?,?,?)")

inputRows = [
    ('aardvark', 1, 0.1),
    ('zymurgy', 2147483647, 99999.999),
    ('foobar', 2000, 9.9)
  ]

for row in inputRows:
   cur.execute(insertStatement,row)
#
# or you can use executemany
#
cur.executemany(insertStatement, inputRows)

Prepared statements are bound to Cursor instance that created them, and can’t be used with any other Cursor instance. Beside repeated execution they are also useful to get information about statement (like its output description, execution plan or statement_type) before its execution.

Example Program:

The following program demonstrates the explicit use of PreparedStatements. It also benchmarks explicit PreparedStatement reuse against normal execution that prepares statements on each execution.

import time
import fdb

con = fdb.connect(dsn='localhost:employee',
    user='sysdba', password='masterkey'
  )

cur = con.cursor()

# Create supporting database entities:
cur.execute("recreate table t (a int, b varchar(50))")
con.commit()
cur.execute("create unique index unique_t_a on t(a)")
con.commit()

# Explicitly prepare the insert statement:
psIns = cur.prep("insert into t (a,b) values (?,?)")
print 'psIns.sql: "%s"' % psIns.sql
print 'psIns.statement_type == fdb.isc_info_sql_stmt_insert:', (
    psIns.statement_type == fdb.isc_info_sql_stmt_insert
  )
print 'psIns.n_input_params: %d' % psIns.n_input_params
print 'psIns.n_output_params: %d' % psIns.n_output_params
print 'psIns.plan: %s' % psIns.plan

print

N = 50000
iStart = 0

# The client programmer uses a PreparedStatement explicitly:
startTime = time.time()
for i in xrange(iStart, iStart + N):
    cur.execute(psIns, (i, str(i)))
print (
    'With explicit prepared statement, performed'
    '\n  %0.2f insertions per second.' % (N / (time.time() - startTime))
  )
con.commit()

iStart += N

# A new SQL string containing the inputs is submitted every time.  Also, in a
# more complicated scenario where the end user supplied the string input
# values, the program would risk SQL injection attacks:
startTime = time.time()
for i in xrange(iStart, iStart + N):
    cur.execute("insert into t (a,b) values (%d,'%s')" % (i, str(i)))
print (
    'When unable to reuse prepared statement, performed'
    '\n  %0.2f insertions per second.' % (N / (time.time() - startTime))
  )
con.commit()

# Prepare a SELECT statement and examine its properties.  The optimizer's plan
# should use the unique index that we created at the beginning of this program.
print
psSel = cur.prep("select * from t where a = ?")
print 'psSel.sql: "%s"' % psSel.sql
print 'psSel.statement_type == fdb.isc_info_sql_stmt_select:', (
    psSel.statement_type == fdb.isc_info_sql_stmt_select
  )
print 'psSel.n_input_params: %d' % psSel.n_input_params
print 'psSel.n_output_params: %d' % psSel.n_output_params
print 'psSel.plan: %s' % psSel.plan

# The current implementation does not allow PreparedStatements to be prepared
# on one Cursor and executed on another:
print
print 'Note that PreparedStatements are not transferrable from one cursor to another:'
cur2 = con.cursor()
cur2.execute(psSel)

Sample output:

psIns.sql: "insert into t (a,b) values (?,?)"
psIns.statement_type == fdb.isc_info_sql_stmt_insert: True
psIns.n_input_params: 2
psIns.n_output_params: 0
psIns.plan: None

With explicit prepared statement, performed
  4276.00 insertions per second.
When unable to reuse prepared statement, performed
  2037.70 insertions per second.

psSel.sql: "select * from t where a = ?"
psSel.statement_type == fdb.isc_info_sql_stmt_select: True
psSel.n_input_params: 1
psSel.n_output_params: 2
psSel.plan: PLAN (T INDEX (UNIQUE_T_A))

Note that PreparedStatements are not transferrable from one cursor to another:
Traceback (most recent call last):
  File "pstest.py", line 85, in <module>
    cur2.execute(psSel)
     File "/home/job/python/envs/pyfirebird/fdb/fdb/fbcore.py", line 2623, in execute
    raise ValueError("PreparedStatement was created by different Cursor.")
ValueError: PreparedStatement was created by different Cursor.

As you can see, the version that prevents the reuse of prepared statements is about two times slower – for a trivial statement. In a real application, SQL statements are likely to be far more complicated, so the speed advantage of using prepared statements would only increase.

Named Cursors¶

To allow the Python programmer to perform scrolling UPDATE or DELETE via the “SELECT … FOR UPDATE” syntax, FDB provides the read/write property Cursor.name.

Example Program:

import fdb

con = fdb.connect(dsn='localhost:/temp/test.db', user='sysdba', password='pass')
curScroll = con.cursor()
curUpdate = con.cursor()

curScroll.execute("select city from addresses for update")
curScroll.name = 'city_scroller'
update = "update addresses set city=? where current of " + curScroll.name

for (city,) in curScroll:
    city = ... # make some changes to city
    curUpdate.execute( update, (city,) )

con.commit()

Working with stored procedures¶

Firebird stored procedures can have input parameters and/or output parameters. Some databases support input/output parameters, where the same parameter is used for both input and output; Firebird does not support this.

It is important to distinguish between procedures that return a result set and procedures that populate and return their output parameters exactly once. Conceptually, the latter “return their output parameters” like a Python function, whereas the former “yield result rows” like a Python generator.

Firebird’s server-side procedural SQL syntax makes no such distinction, but client-side SQL code (and C API code) must. A result set is retrieved from a stored procedure by SELECT’ing from the procedure, whereas output parameters are retrieved with an ‘EXECUTE PROCEDURE’ statement.

To retrieve a result set from a stored procedure with FDB, use code such as this:

cur.execute("select output1, output2 from the_proc(?, ?)", (input1, input2))

# Ordinary fetch code here, such as:
for row in cur:
   ... # process row

con.commit() # If the procedure had any side effects, commit them.

To execute a stored procedure and access its output parameters, use code such as this:

cur.callproc("the_proc", (input1, input2))

# If there are output parameters, retrieve them as though they were the
# first row of a result set.  For example:
outputParams = cur.fetchone()

con.commit() # If the procedure had any side effects, commit them.

This latter is not very elegant; it would be preferable to access the procedure’s output parameters as the return value of Cursor.callproc(). The Python DB API specification requires the current behavior, however.

Implicit Conversion of Input Parameters from Strings¶

The database engine treats most SQL data types in a weakly typed fashion: the engine may attempt to convert the raw value to a different type, as appropriate for the current context. For instance, the SQL expressions 123 (integer) and ‘123’ (string) are treated equivalently when the value is to be inserted into an integer field; the same applies when ‘123’ and 123 are to be inserted into a varchar field.

This weak typing model is quite unlike Python’s dynamic yet strong typing. Although weak typing is regarded with suspicion by most experienced Python programmers, the database engine is in certain situations so aggressive about its typing model that FDB must compromise in order to remain an elegant means of programming the database engine.

An example is the handling of “magic values” for date and time fields. The database engine interprets certain string values such as ‘yesterday’ and ‘now’ as having special meaning in a date/time context. If FDB did not accept strings as the values of parameters destined for storage in date/time fields, the resulting code would be awkward. Consider the difference between the two Python snippets below, which insert a row containing an integer and a timestamp into a table defined with the following DDL statement:

create table test_table (i integer, t timestamp)

i = 1
t = 'now'
sqlWithMagicValues = "insert into test_table (i, t) values (?, '%s')" % t
cur.execute( sqlWithMagicValues, (i,) )

i = 1
t = 'now'
cur.execute( "insert into test_table (i, t) values (?, ?)", (i, t) )

If FDB did not support weak parameter typing, string parameters that the database engine is to interpret as “magic values” would have to be rolled into the SQL statement in a separate operation from the binding of the rest of the parameters, as in the first Python snippet above. Implicit conversion of parameter values from strings allows the consistency evident in the second snippet, which is both more readable and more general.

It should be noted that FDB does not perform the conversion from string itself. Instead, it passes that responsibility to the database engine by changing the parameter metadata structure dynamically at the last moment, then restoring the original state of the metadata structure after the database engine has performed the conversion.

A secondary benefit is that when one uses FDB to import large amounts of data from flat files into the database, the incoming values need not necessarily be converted to their proper Python types before being passed to the database engine. Eliminating this intermediate step may accelerate the import process considerably, although other factors such as the chosen connection protocol and the deactivation of indexes during the import are more consequential. For bulk import tasks, the database engine’s external tables also deserve consideration. External tables can be used to suck semi-structured data from flat files directly into the relational database without the intervention of an ad hoc conversion program.

Automatic conversion from/to unicode¶

In Firebird, every CHAR, VARCHAR or textual BLOB field can (or, better: must) have a character set assigned. While it’s possible to define single character set for whole database, it’s also possible to define different character set for each textual field. This information is used to correctly store the bytes that make up the character string, and together with collation information (that defines the sort ordering and uppercase conversions for a string) is vital for correct data manupulation, including automatic transliteration between character sets when necessary.

FDB helps with client side of this character set bargain by automatically converting unicode strings into bytes/strings encoded in connection character set, and vice versa. However, developers are still responsible that non-unicode strings passed to server are in correct encoding (because FDB makes no assuption about encoding of non-unicode strings, so it can’t recode them to connection charset).

Rules for automatic conversion depend on Python version you use:

• Native Python 2.x strings are passed to server as is, and developers must explicitly use unicode strings to take advantage of automatic conversion. String values coming from server are converted to unicode only:
  • for data stored in database (i.e. not for string values returned by Firebird Service and info calls etc.).
  • when connection charset is specified.
• Native Python 3 strings are unicode, so conversion is fully automatic in both directions for all textual data, i.e. including for string values returned by Firebird Service and info calls etc. When connection charset is not specified, FDB uses locale.getpreferredencoding() to determine encoding for conversions from/to unicode.

Working with BLOBs¶

FDB uses two types of BLOB values:

• Materialized BLOB values are Python strings. This is the default type.
• Streamed BLOB values are file-like objects.

Materialized BLOBs are easy to work with, but are not suitable for:

• deferred loading of BLOBs. They’re called materialized because they’re always fetched from server as part of row fetch. Fetching BLOB value means separate API calls (and network roundtrips), which may slow down you application considerably.
• large values, as they are always stored in memory in full size.

These drawbacks are addressed by stream BLOBs. Using BLOBs in stream mode is easy:

• For input values, simply use parametrized statement and pass any file-like object in place of BLOB parameter. The file-like object must implement only the read() method, as no other metod is used.
• For output values, you have to call Cursor.set_stream_blob() (or PreparedStatement.set_stream_blob()) method with specification of column name(s) that should be returned as file-like objects. FDB then returns BlobReader instance instead string in place of returned BLOB value for these column(s).

The BlobReader instance is bound to particular BLOB value returned by server, so its life time is limited. The actual BLOB value is not opened initially, so no additonal API calls to server are made if you’ll decide to ignore the value completely. You also don’t need to open the BLOB value explicitly, as BLOB is opened automatically on first call to next(), read(), readline(), readlines() or seek(). However, it’s good practice to close() the reader once you’re finished reading, as it’s likely that Python’s garbage collector would call the __del__ method too late, when fetch context is already gone, and closing the reader would cause an error.

Example program:

import os.path
from cStringIO import StringIO

import fdb

con = fdb.connect(dsn='localhost:employee',user='sysdba', password='masterkey')

cur = con.cursor()

cur.execute("recreate table blob_test (a blob)")
con.commit()

# --- Materialized mode (str objects for both input and output) ---
# Insertion:
cur.execute("insert into blob_test values (?)", ('abcdef',))
cur.execute("insert into blob_test values (?)", ('ghijklmnop',))
# Retrieval:
cur.execute("select * from blob_test")
print 'Materialized retrieval (as str):'
print cur.fetchall()

cur.execute("delete from blob_test")

# --- Streaming mode (file-like objects for input; fdb.BlobReader objects for output) ---

# Insertion:
cur.execute("insert into blob_test values (?)", (StringIO('abcdef'),))
cur.execute("insert into blob_test values (?)", (StringIO('ghijklmnop'),))

f = file(os.path.abspath(__file__), 'rb')
cur.execute("insert into blob_test values (?)", (f,))
f.close()

# Retrieval using the "file-like" methods of BlobReader:
cur.execute("select * from blob_test")
cur.set_stream_blob('A') # Note the capital letter

readerA = cur.fetchone()[0]

print '\nStreaming retrieval (via fdb.BlobReader):'

# Python "file-like" interface:
print 'readerA.mode:    "%s"' % readerA.mode
print 'readerA.closed:   %s'  % readerA.closed
print 'readerA.tell():   %d'  % readerA.tell()
print 'readerA.read(2): "%s"' % readerA.read(2)
print 'readerA.tell():   %d'  % readerA.tell()
print 'readerA.read():  "%s"' % readerA.read()
print 'readerA.tell():   %d'  % readerA.tell()
print 'readerA.read():  "%s"' % readerA.read()
readerA.close()
print 'readerA.closed:   %s'  % readerA.closed

Output:

Materialized retrieval (as str):
[('abcdef',), ('ghijklmnop',)]

Streaming retrieval (via fdb.BlobReader):
readerA.mode:    "rb"
readerA.closed:   False
readerA.tell():   0
readerA.read(2): "ab"
readerA.tell():   2
readerA.read():  "cdef"
readerA.tell():   6
readerA.read():  ""
readerA.closed:   True

Firebird ARRAY type¶

FDB supports Firebird ARRAY data type. ARRAY values are represented as Python lists. On input, the Python sequence (list or tuple) must be nested appropriately if the array field is multi-dimensional, and the incoming sequence must not fall short of its maximum possible length (it will not be “padded” implicitly–see below). On output, the lists will be nested if the database array has multiple dimensions.

Example:

>>> import fdb
>>> con = fdb.connect(dsn='localhost:employee',user='sysdba', password='masterkey')
>>> cur = con.cursor()
>>> cur.execute("select LANGUAGE_REQ from job where job_code='Eng' and job_grade=3 and job_country='Japan'")
>>> cur.fetchone()
(['Japanese\n', 'Mandarin\n', 'English\n', '\n', '\n'],)

Example program:

import fdb

con = fdb.connect(dsn='localhost:/temp/test.db', user='sysdba', password='pass')
con.execute_immediate("recreate table array_table (a int[3,4])")
con.commit()

cur = con.cursor()

arrayIn = [
    [1, 2, 3, 4],
    [5, 6, 7, 8],
    [9,10,11,12]
  ]

print 'arrayIn:  %s' % arrayIn
cur.execute("insert into array_table values (?)", (arrayIn,))

cur.execute("select a from array_table")
arrayOut = cur.fetchone()[0]
print 'arrayOut: %s' % arrayOut

con.commit()

Output:

arrayIn:  [[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]]
arrayOut: [[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]]

Basics¶

When it comes to transactions, Python Database API 2.0 specify that Connection object has to respond to the following methods:

Connection.commit()

Commit any pending transaction to the database. Note that if the database supports an auto-commit feature, this must be initially off. An interface method may be provided to turn it back on. Database modules that do not support transactions should implement this method with void functionality.

Connection.rollback()

(optional) In case a database does provide transactions this method causes the the database to roll back to the start of any pending transaction. Closing a connection without committing the changes first will cause an implicit rollback to be performed.

In addition to the implicit transaction initiation required by Python Database API, FDB allows the programmer to start transactions explicitly via the Connection.begin() method. Also Connection.savepoint() method was added to provide support for Firebird SAVEPOINTs.

But Python Database API 2.0 was created with assumption that connection can support only one transactions per single connection. However, Firebird can support multiple independent transactions that can run simultaneously within single connection / attachment to the database. This feature is very important, as applications may require multiple transaction openned simultaneously to perform various tasks, which would require to open multiple connections and thus consume more resources than necessary.

FDB surfaces this Firebird feature by separating transaction management out from Connection into separate Transaction objects. To comply with Python DB API 2.0 requirements, Connection object uses one Transaction instance as main transaction, and delegates begin(), savepoint(), commit(), rollback(), trans_info() and transaction_info() calls to it.

Example:

import fdb

con = fdb.connect(dsn='localhost:employee',user='sysdba', password='masterkey')

cur = con.cursor()

# Most minimalistic transaction management -> implicit start, only commit() and rollback()
# ========================================================================================
#
# Transaction is started implicitly
cur.execute('insert into country values ('Oz','Crowns')
con.commit() # commits active transaction
# Again, transaction is started implicitly
cur.execute('insert into country values ('Barsoom','XXX')
con.rollback() # rolls back active transaction
cur.execute('insert into country values ('Pellucidar','Shells')

# This will roll back the transaction
# because Python DB API 2.0 requires that closing connection
# with pending transaction must cause an implicit rollback
con.close()

Auto-commit¶

FDB doesn’t support auto-commit feature directly, but developers may achieve the similar result using explicit transaction start, taking advantage of default_action and its default value (commit).

Example:

import fdb

con = fdb.connect(dsn='localhost:employee',user='sysdba', password='masterkey')

cur = con.cursor()

con.begin()
cur.execute('insert into country values ('Oz','Crowns')
con.begin() # commits active transaction and starts new one
cur.execute('insert into country values ('Barsoom','XXX')
con.begin() # commits active transaction and starts new one
cur.execute('insert into country values ('Pellucidar','Shells')

# However, commit is required before connection is closed,
# because Python DB API 2.0 requires that closing connection
# with pending transaction must cause an implicit rollback
con.commit()
con.close()

Transaction parameters¶

The database engine offers the client programmer an optional facility called transaction parameter buffers (TPBs) for tweaking the operating characteristics of the transactions he initiates. These include characteristics such as whether the transaction has read and write access to tables, or read-only access, and whether or not other simultaneously active transactions can share table access with the transaction.

Connections have a default_tpb attribute that can be changed to set the default TPB for all transactions subsequently started on the connection. Alternatively, if the programmer only wants to set the TPB for a single transaction, he can start a transaction explicitly via the begin() method and pass a TPB for that single transaction.

For details about TPB construction, see the Firebird API documentation. In particular, the ibase.h supplied with Firebird contains all possible TPB elements – single bytes that the C API defines as constants whose names begin with isc_tpb_. FDB makes all of those TPB constants available (under the same names) as module-level constants. A transaction parameter buffer is handled in C as a character array; FDB requires that TPBs be constructed as Python strings (or bytes for Python 3). Since the constants in the fdb.isc_tpb_* family are numbers, they can’t be simply concatenated to create a TPB, but you may use utility function fdb.bs(byte_array) that accepts sequence of numbers and returns string (P2) or bytes (P3).

For example next call returns TPB for typical READ COMMITED transaction:

from fdb import *

TPB = bs([isc_tpb_version3,
          isc_tpb_write,
          isc_tpb_wait,
          isc_tpb_read_committed,
          isc_tpb_rec_version])

FDB provides several predefined TPB’s for convenience:

• ISOLATION_LEVEL_READ_COMMITED

  Read/Write READ COMMITED with record version and WAIT option. Isolation level with greatest concurrent throughput. This is Default TPB.

  Tip

  This isolation level is optimal for transactions that write data and doesn’t require stable snapshot of database for their operations (i.e. most operations are limited to individual rows).

• ISOLATION_LEVEL_READ_COMMITED_LEGACY

  Read/Write READ COMMITED with NO record version and WAIT option.

  Warning

  This isolation level emulates RDBMS that use locks instead multiversion control (MVC). It’s not recommended to use it at all, except for legacy applications lazily ported from such RDBMS to Firebird.

• ISOLATION_LEVEL_READ_COMMITED_RO

  Like ISOLATION_LEVEL_READ_COMMITED, but Read Only.

  Tip

  Firebird threats these transactions as pre-committed, so they are best option for long running transactions that only read data.

  Internaly FDB uses such transaction to read metadata from connected database. This internal transaction is also available to developers for convenience as Connection.query_transaction.

• ISOLATION_LEVEL_REPEATABLE_READ or ISOLATION_LEVEL_SNAPSHOT

  Read/Write SNAPSHOT (concurrency) with WAIT option.

  Tip

  This isolation level is necessary for transactions that process data in bulk, like reporting, recalculations etc.

• ISOLATION_LEVEL_SERIALIZABLE or ISOLATION_LEVEL_SNAPSHOT_TABLE_STABILITY

  Read/Write SNAPSHOT TABLE STABILITY (consistency) with WAIT option. Like REPEATABLE_READ/SNAPSHOT, but locks whole tables for writes from other transactions. Isolation level with lowest concurrent throughput.

  Warning

  Because tables are locked for protected write (i.e. no other transaction can write until lock is released) at time of first access, there is a great risk of deadlock between transactions.

  Tip

  To prevent deadlocks and increase concurrent throughput it’s recommended to use custom TPB’s with fine-grained table access reservation.

Example:

import fdb

con = fdb.connect(dsn='localhost:employee',user='sysdba', password='masterkey')

cur = con.cursor()

# Start transaction with default_tpb (ISOLATION_LEVEL_READ_COMMITED)
con.begin()
cur.execute('select * from JOB')
com.commit()

# Start using transactions in REPEATABLE READ (SNAPSHOT) isolation
con.default_tpb = fdb.ISOLATION_LEVEL_REPEATABLE_READ
con.begin()
cur.execute('select * from JOB')
com.commit()

# Start THIS transaction as R/O READ COMMITTED
con.begin(fdb.ISOLATION_LEVEL_READ_COMMITED_RO)
cur.execute('select * from JOB')
com.commit()

For cases when predefined transaction parameter blocks are not suitable for your needs, FDB offers utility class TPB for convenient and safe construction of custom tpb blocks. Simply create instance of this class, set member attributes to required values and use either rendered binary tpb block or TPB instance itself to set default_tpb or as paraneter to begin().

Example:

import fdb

con = fdb.connect(dsn='localhost:employee',user='sysdba', password='masterkey')

# Use TPB to construct valid transaction parameter block
# from the fdb.isc_tpb_* family.
customTPB = fdb.TPB()
customTPB.isolation_level = fdb.isc_tpb_consistency # SERIALIZABLE
customTPB.table_reservation["MY_TABLE"] = (fdb.isc_tpb_protected, fdb.isc_tpb_lock_write)

# Explicitly start a transaction with the custom TPB:
con.begin(tpb=customTPB)

# For frequent use, it's better to use already assembled version of TPB
customTPB = fdb.TPB()
customTPB.access_mode = fdb.isc_tpb_read  # read only
customTPB.isolation_level = fdb.isc_tpb_concurrency # SNAPSHOT
customTPB = customTPB.render() # Create valid block according to current values of member attributes.

for x in range(1000):
   con.begin(tpb=customTPB)

If you want to build only table reservation part of tpb (for example to add to various custom built parameter blocks), you can use class TableReservation instead TPB.

Getting information about transaction¶

Transaction object exposes two methods that return information about currently managed active transaction (the same methods are exposed also by Connection object for main_transaction):

transaction_info() is a very thin wrapper around function isc_transaction_info(). This method does not attempt to interpret its results except with regard to whether they are a string or an integer.

A more convenient way to access the same functionality is via the trans_info() method, which is high-level convenience wrapper around the transaction_info method that parses the output of transaction_info into Python-friendly objects instead of returning raw binary buffers in the case of complex result types.

Example program:

import fdb

con = fdb.connect(dsn='localhost:employee',user='sysdba', password='masterkey')

# Start transaction, so we can get information about it
con.begin()

info = con.trans_info([fdb.isc_info_tra_id, fdb.isc_info_tra_oldest_interesting,
                       fdb.isc_info_tra_oldest_snapshot, fdb.isc_info_tra_oldest_active,
                       fdb.isc_info_tra_isolation, fdb.isc_info_tra_access,
                       fdb.isc_info_tra_lock_timeout])

print info
print "TransactionID:", info[fdb.isc_info_tra_id]
print "Oldest Interesting (OIT):",info[fdb.isc_info_tra_oldest_interesting]
print "Oldest Snapshot:",info[fdb.isc_info_tra_oldest_snapshot]
print "Oldest Active (OAT):",info[fdb.isc_info_tra_oldest_active]
print "Isolation Level:",info[fdb.isc_info_tra_isolation]
print "Access Mode:",info[fdb.isc_info_tra_access]
print "Lock Timeout:",info[fdb.isc_info_tra_lock_timeout]

Output:

{4: 459, 5: 430, 6: 459, 7: 459, 8: (3, 1), 9: 1, 10: -1}
TransactionID: 459
Oldest Interesting (OIT): 430
Oldest Snapshot: 459
Oldest Active (OAT): 459
Isolation Level: (3, 1)
Access Mode: 1
Lock Timeout: -1

Retaining transactions¶

The commit() and rollback() methods accept an optional boolean parameter retaining (default False) to indicate whether to recycle the transactional context of the transaction being resolved by the method call.

If retaining is True, the infrastructural support for the transaction active at the time of the method call will be “retained” (efficiently and transparently recycled) after the database server has committed or rolled back the conceptual transaction.

For more information about retaining transactions, see Firebird documentation.

Savepoints¶

Savepoints are named, intermediate control points within an open transaction that can later be rolled back to, without affecting the preceding work. Multiple savepoints can exist within a single unresolved transaction, providing “multi-level undo” functionality.

Although Firebird savepoints are fully supported from SQL alone via the SAVEPOINT ‘name’ and ROLLBACK TO ‘name’ statements, FDB also exposes savepoints at the Python API level for the sake of convenience.

Call to method savepoint() establishes a savepoint with the specified name. To roll back to a specific savepoint, call the rollback() method and provide a value (the name of the savepoint) for the optional savepoint parameter. If the savepoint parameter of rollback() is not specified, the active transaction is cancelled in its entirety, as required by the Python Database API Specification.

The following program demonstrates savepoint manipulation via the FDB API, rather than raw SQL.

import fdb

con = fdb.connect(dsn='employee', user='sysdba', password='pass')
cur = con.cursor()

cur.execute("recreate table test_savepoints (a integer)")
con.commit()

print 'Before the first savepoint, the contents of the table are:'
cur.execute("select * from test_savepoints")
print ' ', cur.fetchall()

cur.execute("insert into test_savepoints values (?)", [1])
con.savepoint('A')
print 'After savepoint A, the contents of the table are:'
cur.execute("select * from test_savepoints")
print ' ', cur.fetchall()

cur.execute("insert into test_savepoints values (?)", [2])
con.savepoint('B')
print 'After savepoint B, the contents of the table are:'
cur.execute("select * from test_savepoints")
print ' ', cur.fetchall()

cur.execute("insert into test_savepoints values (?)", [3])
con.savepoint('C')
print 'After savepoint C, the contents of the table are:'
cur.execute("select * from test_savepoints")
print ' ', cur.fetchall()

con.rollback(savepoint='A')
print 'After rolling back to savepoint A, the contents of the table are:'
cur.execute("select * from test_savepoints")
print ' ', cur.fetchall()

con.rollback()
print 'After rolling back entirely, the contents of the table are:'
cur.execute("select * from test_savepoints")
print ' ', cur.fetchall()

The output of the example program is shown below:

Before the first savepoint, the contents of the table are:
  []
After savepoint A, the contents of the table are:
  [(1,)]
After savepoint B, the contents of the table are:
  [(1,), (2,)]
After savepoint C, the contents of the table are:
  [(1,), (2,), (3,)]
After rolling back to savepoint A, the contents of the table are:
  [(1,)]
After rolling back entirely, the contents of the table are:
  []

Using multiple transactions with the same connection¶

To use additional transactions that could run simultaneously with main transaction managed by Connection, create new Transaction object calling Connection.trans() method. If you don’t specify the optional default_tpb parameter, this new Transaction inherits the default_tpb from Connection. Physical transaction is not started when Transaction instance is created, but implicitly when first SQL statement is executed, or explicitly via Transaction.begin() call.

To execute statements in context of this additional transaction you have to use cursors obtained directly from this Transaction instance calling its cursor() method, or call Transaction.execute_immediate() method.

Example:

import fdb

con = fdb.connect(dsn='employee', user='sysdba', password='pass')
# Cursor for main_transaction context
cur = con.cursor()

# Create new READ ONLY READ COMMITTED transaction
ro_transaction = con.trans(fdb.ISOLATION_LEVEL_READ_COMMITED_RO)
# and cursor
ro_cur = ro_transaction.cursor()

cur.execute('insert into country values ('Oz','Crowns')
con.commit() # commits main transaction

# Read data created by main transaction from second one
ro_cur.execute("select * from COUNTRY where COUNTRY = `Oz`")
print ro_cur.fetchall()

# Insert more data, but don't commit
cur.execute('insert into country values ('Barsoom','XXX')

# Read data created by main transaction from second one
ro_cur.execute("select * from COUNTRY where COUNTRY = `Barsoom`")
print ro_cur.fetchall()

Distributed Transactions¶

Distributed transactions are transactions that span multiple databases. FDB provides this Firebird feature through ConnectionGroup class. Instances of this class act as managers for Transaction object that is bound to multiple connections, and to cursors bound to it and connections participated in group. That’s it, distributed transaction is fully independent from all other transactions, main or secondary, of member connections.

To assemble a group of connections, you can either pass the sequence of Connection instances to ConnectionGroup constructor, or add connections latter calling ConnectionGroup.add() method.

Any Connection could be a member of only one group, and attempt to add it to another one would raise an exception. Also, Connection participating in group cannot be closed before it’s removed or whole group is disbanded.

Similarly to Transaction, distributed transactions are managed through ConnectionGroup.begin(), ConnectionGroup.savepoint(). ConnectionGroup.commit() and ConnectionGroup.rollback() methods. Additionally, ConnectionGroup exposes method prepare() that explicitly initiates the first phase of Two-Phase Commit Protocol. Transaction parameters are defined similarly to Transaction using ConnectionGroup.default_tpb or as optional parameter to begin() call.

SQL statements that should belong to context of distributed transaction are executed via Cursor instances aquired through ConnectionGroup.cursor() method, or calling ConnectionGroup.execute_immediate() method.

Example program:

import fdb

# First database
con1 = fdb.create_database("CREATE DATABASE 'testdb-1.fdb' USER 'SYSDBA' PASSWORD 'masterkey'")
con1.execute_immediate("recreate table T (PK integer, C1 integer)")
con1.commit()

# Second database
con2 = fdb.create_database("CREATE DATABASE 'testdb-2,fdb' USER 'SYSDBA' PASSWORD 'masterkey'")
con2.execute_immediate("recreate table T (PK integer, C1 integer)")
con2.commit()

# Create connection group
cg = fdb.ConnectionGroup((con1,con2))

# Prepare Group cursors for each connection
gc1 = cg.cursor(con1)
gc2 = cg.cursor(con2)

# Connection cursors to check content of databases
q = 'select * from T order by pk'

cc1 = con1.cursor()
p1 = cc1.prep(q)

cc2 = con2.cursor()
p2 = cc2.prep(q)

print "Distributed transaction: COMMIT"
#      ===============================
gc1.execute('insert into t (pk) values (1)')
gc2.execute('insert into t (pk) values (1)')
cg.commit()

# check it
con1.commit()
cc1.execute(p1)
print 'db1:',cc1.fetchall()
con2.commit()
cc2.execute(p2)
print 'db2:',cc2.fetchall()

print "Distributed transaction: PREPARE + COMMIT"
#      =========================================
gc1.execute('insert into t (pk) values (2)')
gc2.execute('insert into t (pk) values (2)')
cg.prepare()
cg.commit()

# check it
con1.commit()
cc1.execute(p1)
print 'db1:',cc1.fetchall()
con2.commit()
cc2.execute(p2)
print 'db2:',cc2.fetchall()

print "Distributed transaction: SAVEPOINT + ROLLBACK to it"
#      ===================================================
gc1.execute('insert into t (pk) values (3)')
cg.savepoint('CG_SAVEPOINT')
gc2.execute('insert into t (pk) values (3)')
cg.rollback(savepoint='CG_SAVEPOINT')

# check it - via group cursors, as transaction is still active
gc1.execute(q)
print 'db1:',gc1.fetchall()
gc2.execute(q)
print 'db2:',gc2.fetchall()

print "Distributed transaction: ROLLBACK"
#      =================================
cg.rollback()

# check it
con1.commit()
cc1.execute(p1)
print 'db1:',cc1.fetchall()
con2.commit()
cc2.execute(p2)
print 'db2:',cc2.fetchall()

print "Distributed transaction: EXECUTE_IMMEDIATE"
#      ==========================================
cg.execute_immediate('insert into t (pk) values (3)')
cg.commit()

# check it
con1.commit()
cc1.execute(p1)
print 'db1:',cc1.fetchall()
con2.commit()
cc2.execute(p2)
print 'db2:',cc2.fetchall()

# Finalize
con1.drop_database()
con1.close()
con2.drop_database()
con2.close()

Output:

Distributed transaction: COMMIT
db1: [(1, None)]
db2: [(1, None)]
Distributed transaction: PREPARE + COMMIT
db1: [(1, None), (2, None)]
db2: [(1, None), (2, None)]
Distributed transaction: SAVEPOINT + ROLLBACK to it
db1: [(1, None), (2, None), (3, None)]
db2: [(1, None), (2, None)]
Distributed transaction: ROLLBACK
db1: [(1, None), (2, None)]
db2: [(1, None), (2, None)]
Distributed transaction: EXECUTE_IMMEDIATE
db1: [(1, None), (2, None), (3, None)]
db2: [(1, None), (2, None), (3, None)]

Transaction Context Manager¶

FDB provides context manager TransactionContext that allows automatic transaction management using The with statement. It can work with any object that supports begin(), commit() and rollback() methods, i.e. Connection, ConnectionGroup or Transaction.

It starts transaction when WITH block is entered and commits it if no exception occurst within it, or calls rollback() otherwise. Exceptions raised in WITH block are never suppressed.

Examples:

con = fdb.connect(dsn='employee',user='sysdba',password='masterkey')

# Uses default main transaction
with TransactionContext(con):
   cur = con.cursor()
   cur.execute("insert into T (PK,C1) values (1,'TXT')")

# Uses separate transaction
with TransactionContext(con.trans()) as tr:
   cur = tr.cursor()
   cur.execute("insert into T (PK,C1) values (2,'AAA')")

# Uses connection group (distributed transaction)
con2 = fdb.connect(dsn='remote:employee',user='sysdba',password='masterkey')
cg = fdb.ConnectionGroup((con,con2))
with TransactionContext(cg):
   cur1 = cg.cursor(con)
   cur2 = cg.cursor(con2)
   cur1.execute("insert into T (PK,C1) values (3,'Local')")
   cur2.execute("insert into T (PK,C1) values (3,'Remote')")

What they are¶

The Firebird engine features a distributed, interprocess communication mechanism based on messages called database events. A database event is a message passed from a trigger or stored procedure to an application to announce the occurrence of a specified condition or action, usually a database change such as an insertion, modification, or deletion of a record. The Firebird event mechanism enables applications to respond to actions and database changes made by other, concurrently running applications without the need for those applications to communicate directly with one another, and without incurring the expense of CPU time required for periodic polling to determine if an event has occurred.

Why use them¶

Anything that can be accomplished with database events can also be implemented using other techniques, so why bother with events? Since you’ve chosen to write database-centric programs in Python rather than assembly language, you probably already know the answer to this question, but let’s illustrate.

A typical application for database events is the handling of administrative messages. Suppose you have an administrative message database with a message’s table, into which various applications insert timestamped status reports. It may be desirable to react to these messages in diverse ways, depending on the status they indicate: to ignore them, to initiate the update of dependent databases upon their arrival, to forward them by e-mail to a remote administrator, or even to set off an alarm so that on-site administrators will know a problem has occurred.

It is undesirable to tightly couple the program whose status is being reported (the message producer) to the program that handles the status reports (the message handler). There are obvious losses of flexibility in doing so. For example, the message producer may run on a separate machine from the administrative message database and may lack access rights to the downstream reporting facilities (e.g., network access to the SMTP server, in the case of forwarded e-mail notifications). Additionally, the actions required to handle status reports may themselves be time-consuming and error-prone, as in accessing a remote network to transmit e-mail.

In the absence of database event support, the message handler would probably be implemented via polling. Polling is simply the repetition of a check for a condition at a specified interval. In this case, the message handler would check in an infinite loop to see whether the most recent record in the messages table was more recent than the last message it had handled. If so, it would handle the fresh message(s); if not, it would go to sleep for a specified interval, then loop.

The polling-based implementation of the message handler is fundamentally flawed. Polling is a form of busy-wait; the check for new messages is performed at the specified interval, regardless of the actual activity level of the message producers. If the polling interval is lengthy, messages might not be handled within a reasonable time period after their arrival; if the polling interval is brief, the message handler program (and there may be many such programs) will waste a large amount of CPU time on unnecessary checks.

The database server is necessarily aware of the exact moment when a new message arrives. Why not let the message handler program request that the database server send it a notification when a new message arrives? The message handler can then efficiently sleep until the moment its services are needed. Under this event-based scheme, the message handler becomes aware of new messages at the instant they arrive, yet it does not waste CPU time checking in vain for new messages when there are none available.

How events are exposed¶

1. Server Process (“An event just occurred!”)

   To notify any interested listeners that a specific event has occurred, issue the POST_EVENT statement from Stored Procedure or Trigger. The POST_EVENT statement has one parameter: the name of the event to post. In the preceding example of the administrative message database, POST_EVENT might be used from an after insert trigger on the messages table, like this:

   create trigger trig_messages_handle_insert
     for messages
       after insert
   as
   begin
     POST_EVENT 'new_message';
   end
   

   Note

   The physical notification of the client process does not occur until the transaction in which the POST_EVENT took place is actually committed. Therefore, multiple events may conceptually occur before the client process is physically informed of even one occurrence. Furthermore, the database engine makes no guarantee that clients will be informed of events in the same groupings in which they conceptually occurred. If, within a single transaction, an event named event_a is posted once and an event named event_b is posted once, the client may receive those posts in separate “batches”, despite the fact that they occurred in the same conceptual unit (a single transaction). This also applies to multiple occurrences of the same event within a single conceptual unit: the physical notifications may arrive at the client separately.

2. Client Process (“Send me a message when an event occurs.”)

   Note

   If you don’t care about the gory details of event notification, skip to the section that describes FDB’s Python-level event handling API.

   The Firebird C client library offers two forms of event notification. The first form is synchronous notification, by way of the function isc_wait_for_event(). This form is admirably simple for a C programmer to use, but is inappropriate as a basis for FDB’s event support, chiefly because it’s not sophisticated enough to serve as the basis for a comfortable Python-level API. The other form of event notification offered by the database client library is asynchronous, by way of the functions isc_que_events() (note that the name of that function is misspelled), isc_cancel_events(), and others. The details are as nasty as they are numerous, but the essence of using asynchronous notification from C is as follows:

   1. Call isc_event_block() to create a formatted binary buffer that will tell the server which events the client wants to listen for.
   2. Call isc_que_events() (passing the buffer created in the previous step) to inform the server that the client is ready to receive event notifications, and provide a callback that will be asynchronously invoked when one or more of the registered events occurs.
   3. [The thread that called isc_que_events() to initiate event listening must now do something else.]
   4. When the callback is invoked (the database client library starts a thread dedicated to this purpose), it can use the isc_event_counts() function to determine how many times each of the registered events has occurred since the last call to isc_event_counts() (if any).
   5. [The callback thread should now “do its thing”, which may include communicating with the thread that called isc_que_events().]
   6. When the callback thread is finished handling an event notification, it must call isc_que_events() again in order to receive future notifications. Future notifications will invoke the callback again, effectively “looping” the callback thread back to Step 4.

API for Python developers¶

The FDB database event API is comprised of the following: the method Connection.event_conduit() and the class EventConduit.

The EventConduit class serve as “conduit” through which database event notifications will flow into the Python program. It’s not designed to be instantiated directly by the Python programmer. Instead, use the Connection.event_conduit() method to create EventConduit instances. event_conduit is a method of Connection rather than a module-level function or a class constructor because the database engine deals with events in the context of a particular database (after all, POST_EVENT must be issued by a stored procedure or a trigger).

Connection.event_conduit() takes a sequence of string event names as parameter, and returns EventConduit instance.

Immediately when begin() method is called, EventConduit starts to accumulate notifications of any events that occur within the conduit’s internal queue until the conduit is closed either explicitly (via the close() method) or implicitly (via garbage collection).

Notifications about events are aquired through call to wait() method, that blocks the calling thread until at least one of the events occurs, or the specified timeout (if any) expires, and returns None if the wait timed out, or a dictionary that maps event_name -> event_occurrence_count.

Example:

with connection.event_conduit( ('event_a', 'event_b') ) as conduit:
    events = conduit.wait()
    process_events(events)

If you want to drop notifications accumulated so far by conduit, call EventConduit.flush() method.

Example program:

import fdb
import threading
import time

# Prepare database
con = fdb.create_database("CREATE DATABASE 'event_test.fdb' USER 'SYSDBA' PASSWORD 'masterkey'")
con.execute_immediate("CREATE TABLE T (PK Integer, C1 Integer)")
con.execute_immediate("""CREATE TRIGGER EVENTS_AU FOR T ACTIVE
BEFORE UPDATE POSITION 0
AS
BEGIN
   if (old.C1 <> new.C1) then
      post_event 'c1_updated' ;
END""")
con.execute_immediate("""CREATE TRIGGER EVENTS_AI FOR T ACTIVE
AFTER INSERT POSITION 0
AS
BEGIN
   if (new.c1 = 1) then
     post_event 'insert_1' ;
   else if (new.c1 = 2) then
     post_event 'insert_2' ;
   else if (new.c1 = 3) then
     post_event 'insert_3' ;
   else
     post_event 'insert_other' ;
END""")
con.commit()
cur = con.cursor()

# Utility function
def send_events(command_list):
   for cmd in command_list:
      cur.execute(cmd)
   con.commit()

print "One event"
#      =========
timed_event = threading.Timer(3.0,send_events,args=[["insert into T (PK,C1) values (1,1)",]])
events = con.event_conduit(['insert_1'])
events.begin()
timed_event.start()
e = events.wait()
events.close()
print e

print "Multiple events"
#      ===============
cmds = ["insert into T (PK,C1) values (1,1)",
        "insert into T (PK,C1) values (1,2)",
        "insert into T (PK,C1) values (1,3)",
        "insert into T (PK,C1) values (1,1)",
        "insert into T (PK,C1) values (1,2)",]
timed_event = threading.Timer(3.0,send_events,args=[cmds])
events = self.con.event_conduit(['insert_1','insert_3'])
events.begin()
timed_event.start()
e = events.wait()
events.close()
print e

print "20 events"
#      =========
cmds = ["insert into T (PK,C1) values (1,1)",
        "insert into T (PK,C1) values (1,2)",
        "insert into T (PK,C1) values (1,3)",
        "insert into T (PK,C1) values (1,1)",
        "insert into T (PK,C1) values (1,2)",]
timed_event = threading.Timer(1.0,send_events,args=[cmds])
events = con.event_conduit(['insert_1','A','B','C','D',
                            'E','F','G','H','I','J','K','L','M',
                            'N','O','P','Q','R','insert_3'])
events.begin()
timed_event.start()
time.sleep(3)
e = events.wait()
events.close()
print e

print "Flush events"
#      ============
timed_event = threading.Timer(3.0,send_events,args=[["insert into T (PK,C1) values (1,1)",]])
events = con.event_conduit(['insert_1'])
events.begin()
send_events(["insert into T (PK,C1) values (1,1)",
             "insert into T (PK,C1) values (1,1)"])
time.sleep(2)
events.flush()
timed_event.start()
e = events.wait()
events.close()
print e

# Finalize
con.drop_database()
con.close()

Output:

One event
{'insert_1': 1}
Multiple events
{'insert_3': 1, 'insert_1': 2}
20 events
{'A': 0, 'C': 0, 'B': 0, 'E': 0, 'D': 0, 'G': 0, 'insert_1': 2, 'I': 0, 'H': 0, 'K': 0, 'J': 0, 'M': 0,
 'L': 0, 'O': 0, 'N': 0, 'Q': 0, 'P': 0, 'R': 0, 'insert_3': 1, 'F': 0}
Flush events
{'insert_1': 1}

Services API Connections¶

All Services API operations are performed in the context of a connection to a specific database server, represented by the fdb.services.Connection class. Similarly to database connections, FDB provides connect() constructor function to create such connections.

This constructor has three keyword parameters:

Since maintenance operations are most often initiated by an administrative user on the same computer as the database server, host defaults to the local computer, and user defaults to SYSDBA.

The three calls to fdb.services.connect() in the following program are equivalent:

from fdb import services

con = services.connect(password='masterkey')
con = services.connect(user='sysdba', password='masterkey')
con = services.connect(host='localhost', user='sysdba', password='masterkey')

Connection object provides number of methods that could be divided into several groups:

• Server Configuration and State: To get information about server configuration, active attachments or users, or to get content of server log.
• Database options: To set various database parameters like size of page cache, access mode or SQL dialect.
• Database maintenance: To perform backup, restore, validation or other database maintenance tasks.
• User maintenance: To get or change information about users defined in security database, to create new or remove users.
• Trace service: To start, stop, pause/resume or list Firebird trace sessions.
• Text ouput from Services: Some services like backup or trace may return significant amount of text. This output is not returned directly by method that starts the service, but through separate methods that emulate read from text file, or provide iterator protocol support on Connection.

Server Configuration and State¶

get_service_manager_version()

To help client programs adapt to version changes, the service manager exposes its version number as an integer.

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> print con.get_service_manager_version()
2

fdb.services is a thick wrapper of the Services API that can shield its users from changes in the underlying C API, so this method is unlikely to be useful to the typical Python client programmer.

get_server_version()

Returns the server’s version string

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> print con.get_server_version()
LI-V2.5.2.26536 Firebird 2.5

At first glance, this method appears to duplicate the functionality of the fdb.Connection.server_version property, but when working with Firebird, there is a difference. fdb.Connection.server_version is based on a C API call (isc_database_info()) that existed long before the introduction of the Services API. Some programs written before the advent of Firebird test the version number in the return value of isc_database_info(), and refuse to work if it indicates that the server is too old. Since the first stable version of Firebird was labeled 1.0, this pre-Firebird version testing scheme incorrectly concludes that (e.g.) Firebird 1.0 is older than Interbase 5.0.

Firebird addresses this problem by making isc_database_info() return a “pseudo-InterBase” version number, whereas the Services API returns the true Firebird version, as shown:

# 64-bit Linux Firebird 2.5.1 SuperServer
import fdb
con = fdb.connect(dsn='employee', user='sysdba', password='masterkey')
print 'Interbase-compatible version string:', con.server_version
svcCon = fdb.services.connect(password='masterkey')
print 'Actual Firebird version string:     ', svcCon.get_server_version()

Output (on Firebird 2.5.1/Linux64):

Interbase-compatible version string: LI-V6.3.1.26351 Firebird 2.5
Actual Firebird version string:      LI-V2.5.1.26351 Firebird 2.5

get_architecture()

Returns platform information for the server, including hardware architecture and operating system family.

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> print con.get_architecture()
Firebird/linux AMD64

get_home_directory()

Returns the equivalent of the RootDirectory setting from firebird.conf.

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> print con.get_home_directory()
/opt/firebird/

get_security_database_path()

Returns the location of the server’s core security database, which contains user definitions and such. Name of this database is security2.fdb (Firebird 2.0 and later) or security.fdb (Firebird 1.5).

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> print con.get_security_database_path()
/opt/firebird/security2.fdb

get_lock_file_directory()

Returns the directory location for Firebird lock files.

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> print con.get_lock_file_directory()
/tmp/firebird/

get_server_capabilities()

Returns tuple of capability info codes for each capability reported by Firebird server. Following constants are defined in fdb.services for convenience:

• CAPABILITY_MULTI_CLIENT
• CAPABILITY_REMOTE_HOP
• CAPABILITY_SERVER_CONFIG
• CAPABILITY_QUOTED_FILENAME
• CAPABILITY_NO_SERVER_SHUTDOWN

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> print con.get_server_capabilities()
(2L, 4L, 512L, 256L)
>>> fdb.services.CAPABILITY_MULTI_CLIENT in con.get_server_capabilities()
True
>>> fdb.services.CAPABILITY_QUOTED_FILENAME in con.get_server_capabilities()
False

get_message_file_directory()

To support internationalized error messages/prompts, the database engine stores its messages in a file named firebird.msg. The directory in which this file resides can be determined with this method.

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> print con.get_message_file_directory()
/opt/firebird/

get_connection_count()

Returns the number of active connections to databases managed by the server. This count only includes database connections (such as open instances of fdb.Connection), not services manager connections (such as open instances of fdb.services.Connection).

# 64-bit Linux Firebird 2.5.1 SuperServer
>>> from fdb import services
>>> con = services.connect(host='localhost', user='sysdba', password='masterkey')
>>> db1 = fdb.connect(dsn='employee',user='sysdba',password='masterkey')
>>> db2 = fdb.connect(dsn='employee',user='sysdba',password='masterkey')
>>> print con.get_connection_count()
2