django1/django/db/backends/postgresql/operations.py

95 lines
4.6 KiB
Python
Raw Normal View History

from django.db.backends import BaseDatabaseOperations
# This DatabaseOperations class lives in here instead of base.py because it's
# used by both the 'postgresql' and 'postgresql_psycopg2' backends.
class DatabaseOperations(BaseDatabaseOperations):
def __init__(self, postgres_version=None):
self.postgres_version = postgres_version
def date_extract_sql(self, lookup_type, field_name):
# http://www.postgresql.org/docs/8.0/static/functions-datetime.html#FUNCTIONS-DATETIME-EXTRACT
return "EXTRACT('%s' FROM %s)" % (lookup_type, field_name)
def date_trunc_sql(self, lookup_type, field_name):
# http://www.postgresql.org/docs/8.0/static/functions-datetime.html#FUNCTIONS-DATETIME-TRUNC
return "DATE_TRUNC('%s', %s)" % (lookup_type, field_name)
def deferrable_sql(self):
return " DEFERRABLE INITIALLY DEFERRED"
def last_insert_id(self, cursor, table_name, pk_name):
cursor.execute("SELECT CURRVAL('\"%s_%s_seq\"')" % (table_name, pk_name))
return cursor.fetchone()[0]
def quote_name(self, name):
if name.startswith('"') and name.endswith('"'):
return name # Quoting once is enough.
return '"%s"' % name
def sql_flush(self, style, tables, sequences):
if tables:
if self.postgres_version[0] >= 8 and self.postgres_version[1] >= 1:
# Postgres 8.1+ can do 'TRUNCATE x, y, z...;'. In fact, it *has to*
# in order to be able to truncate tables referenced by a foreign
# key in any other table. The result is a single SQL TRUNCATE
# statement.
sql = ['%s %s;' % \
(style.SQL_KEYWORD('TRUNCATE'),
style.SQL_FIELD(', '.join([self.quote_name(table) for table in tables]))
)]
else:
# Older versions of Postgres can't do TRUNCATE in a single call, so
# they must use a simple delete.
sql = ['%s %s %s;' % \
(style.SQL_KEYWORD('DELETE'),
style.SQL_KEYWORD('FROM'),
style.SQL_FIELD(self.quote_name(table))
) for table in tables]
# 'ALTER SEQUENCE sequence_name RESTART WITH 1;'... style SQL statements
# to reset sequence indices
for sequence_info in sequences:
table_name = sequence_info['table']
column_name = sequence_info['column']
if column_name and len(column_name) > 0:
sequence_name = '%s_%s_seq' % (table_name, column_name)
else:
sequence_name = '%s_id_seq' % table_name
sql.append("%s setval('%s', 1, false);" % \
(style.SQL_KEYWORD('SELECT'),
style.SQL_FIELD(self.quote_name(sequence_name)))
)
return sql
else:
return []
def sequence_reset_sql(self, style, model_list):
from django.db import models
output = []
qn = self.quote_name
for model in model_list:
# Use `coalesce` to set the sequence for each model to the max pk value if there are records,
# or 1 if there are none. Set the `is_called` property (the third argument to `setval`) to true
# if there are records (as the max pk value is already in use), otherwise set it to false.
for f in model._meta.fields:
if isinstance(f, models.AutoField):
output.append("%s setval('%s', coalesce(max(%s), 1), max(%s) %s null) %s %s;" % \
(style.SQL_KEYWORD('SELECT'),
style.SQL_FIELD(qn('%s_%s_seq' % (model._meta.db_table, f.column))),
style.SQL_FIELD(qn(f.column)),
style.SQL_FIELD(qn(f.column)),
style.SQL_KEYWORD('IS NOT'),
style.SQL_KEYWORD('FROM'),
style.SQL_TABLE(qn(model._meta.db_table))))
break # Only one AutoField is allowed per model, so don't bother continuing.
for f in model._meta.many_to_many:
output.append("%s setval('%s', coalesce(max(%s), 1), max(%s) %s null) %s %s;" % \
(style.SQL_KEYWORD('SELECT'),
style.SQL_FIELD(qn('%s_id_seq' % f.m2m_db_table())),
style.SQL_FIELD(qn('id')),
style.SQL_FIELD(qn('id')),
style.SQL_KEYWORD('IS NOT'),
style.SQL_KEYWORD('FROM'),
style.SQL_TABLE(f.m2m_db_table())))
return output