/usr/lib/python2.7/dist-packages/sqlobject/sresults.py is in python-sqlobject 1.5.2-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 | import dbconnection
import joins
import main
import sqlbuilder
__all__ = ['SelectResults']
class SelectResults(object):
IterationClass = dbconnection.Iteration
def __init__(self, sourceClass, clause, clauseTables=None,
**ops):
self.sourceClass = sourceClass
if clause is None or isinstance(clause, str) and clause == 'all':
clause = sqlbuilder.SQLTrueClause
if not isinstance(clause, sqlbuilder.SQLExpression):
clause = sqlbuilder.SQLConstant(clause)
self.clause = clause
self.ops = ops
if ops.get('orderBy', sqlbuilder.NoDefault) is sqlbuilder.NoDefault:
ops['orderBy'] = sourceClass.sqlmeta.defaultOrder
orderBy = ops['orderBy']
if isinstance(orderBy, (tuple, list)):
orderBy = map(self._mungeOrderBy, orderBy)
else:
orderBy = self._mungeOrderBy(orderBy)
ops['dbOrderBy'] = orderBy
if 'connection' in ops and ops['connection'] is None:
del ops['connection']
if ops.get('limit', None):
assert not ops.get('start', None) and not ops.get('end', None), \
"'limit' cannot be used with 'start' or 'end'"
ops["start"] = 0
ops["end"] = ops.pop("limit")
tablesSet = sqlbuilder.tablesUsedSet(self.clause, self._getConnection().dbName)
if clauseTables:
for table in clauseTables:
tablesSet.add(table)
self.clauseTables = clauseTables
# Explicitly post-adding-in sqlmeta.table, sqlbuilder.Select will handle sqlrepr'ing and dupes
self.tables = list(tablesSet) + [sourceClass.sqlmeta.table]
def queryForSelect(self):
columns = [self.sourceClass.q.id] + [getattr(self.sourceClass.q, x.name) for x in self.sourceClass.sqlmeta.columnList]
query = sqlbuilder.Select(columns,
where=self.clause,
join=self.ops.get('join', sqlbuilder.NoDefault),
distinct=self.ops.get('distinct',False),
lazyColumns=self.ops.get('lazyColumns', False),
start=self.ops.get('start', 0),
end=self.ops.get('end', None),
orderBy=self.ops.get('dbOrderBy',sqlbuilder.NoDefault),
reversed=self.ops.get('reversed', False),
staticTables=self.tables,
forUpdate=self.ops.get('forUpdate', False))
return query
def __repr__(self):
return "<%s at %x>" % (self.__class__.__name__, id(self))
def _getConnection(self):
return self.ops.get('connection') or self.sourceClass._connection
def __str__(self):
conn = self._getConnection()
return conn.queryForSelect(self)
def _mungeOrderBy(self, orderBy):
if isinstance(orderBy, str) and orderBy.startswith('-'):
orderBy = orderBy[1:]
desc = True
else:
desc = False
if isinstance(orderBy, basestring):
if orderBy in self.sourceClass.sqlmeta.columns:
val = getattr(self.sourceClass.q, self.sourceClass.sqlmeta.columns[orderBy].name)
if desc:
return sqlbuilder.DESC(val)
else:
return val
else:
orderBy = sqlbuilder.SQLConstant(orderBy)
if desc:
return sqlbuilder.DESC(orderBy)
else:
return orderBy
else:
return orderBy
def clone(self, **newOps):
ops = self.ops.copy()
ops.update(newOps)
return self.__class__(self.sourceClass, self.clause,
self.clauseTables, **ops)
def orderBy(self, orderBy):
return self.clone(orderBy=orderBy)
def connection(self, conn):
return self.clone(connection=conn)
def limit(self, limit):
return self[:limit]
def lazyColumns(self, value):
return self.clone(lazyColumns=value)
def reversed(self):
return self.clone(reversed=not self.ops.get('reversed', False))
def distinct(self):
return self.clone(distinct=True)
def newClause(self, new_clause):
return self.__class__(self.sourceClass, new_clause,
self.clauseTables, **self.ops)
def filter(self, filter_clause):
if filter_clause is None:
# None doesn't filter anything, it's just a no-op:
return self
clause = self.clause
if isinstance(clause, basestring):
clause = sqlbuilder.SQLConstant('(%s)' % clause)
return self.newClause(sqlbuilder.AND(clause, filter_clause))
def __getitem__(self, value):
if isinstance(value, slice):
assert not value.step, "Slices do not support steps"
if not value.start and not value.stop:
# No need to copy, I'm immutable
return self
# Negative indexes aren't handled (and everything we
# don't handle ourselves we just create a list to
# handle)
if (value.start and value.start < 0) \
or (value.stop and value.stop < 0):
if value.start:
if value.stop:
return list(self)[value.start:value.stop]
return list(self)[value.start:]
return list(self)[:value.stop]
if value.start:
assert value.start >= 0
start = self.ops.get('start', 0) + value.start
if value.stop is not None:
assert value.stop >= 0
if value.stop < value.start:
# an empty result:
end = start
else:
end = value.stop + self.ops.get('start', 0)
if self.ops.get('end', None) is not None and \
self.ops['end'] < end:
# truncated by previous slice:
end = self.ops['end']
else:
end = self.ops.get('end', None)
else:
start = self.ops.get('start', 0)
end = value.stop + start
if self.ops.get('end', None) is not None \
and self.ops['end'] < end:
end = self.ops['end']
return self.clone(start=start, end=end)
else:
if value < 0:
return list(iter(self))[value]
else:
start = self.ops.get('start', 0) + value
return list(self.clone(start=start, end=start+1))[0]
def __iter__(self):
# @@: This could be optimized, using a simpler algorithm
# since we don't have to worry about garbage collection,
# etc., like we do with .lazyIter()
return iter(list(self.lazyIter()))
def lazyIter(self):
"""
Returns an iterator that will lazily pull rows out of the
database and return SQLObject instances
"""
conn = self._getConnection()
return conn.iterSelect(self)
def accumulate(self, *expressions):
""" Use accumulate expression(s) to select result
using another SQL select through current
connection.
Return the accumulate result
"""
conn = self._getConnection()
exprs = []
for expr in expressions:
if not isinstance(expr, sqlbuilder.SQLExpression):
expr = sqlbuilder.SQLConstant(expr)
exprs.append(expr)
return conn.accumulateSelect(self, *exprs)
def count(self):
""" Counting elements of current select results """
assert not self.ops.get('start') and not self.ops.get('end'), \
"start/end/limit have no meaning with 'count'"
assert not (self.ops.get('distinct') and (self.ops.get('start')
or self.ops.get('end'))), \
"distinct-counting of sliced objects is not supported"
if self.ops.get('distinct'):
# Column must be specified, so we are using unique ID column.
# COUNT(DISTINCT column) is supported by MySQL and PostgreSQL,
# but not by SQLite. Perhaps more portable would be subquery:
# SELECT COUNT(*) FROM (SELECT DISTINCT id FROM table)
count = self.accumulate('COUNT(DISTINCT %s)' % self._getConnection().sqlrepr(self.sourceClass.q.id))
else:
count = self.accumulate('COUNT(*)')
if self.ops.get('start'):
count -= self.ops['start']
if self.ops.get('end'):
count = min(self.ops['end'] - self.ops.get('start', 0), count)
return count
def accumulateMany(self, *attributes):
""" Making the expressions for count/sum/min/max/avg
of a given select result attributes.
`attributes` must be a list/tuple of pairs (func_name, attribute);
`attribute` can be a column name (like 'a_column')
or a dot-q attribute (like Table.q.aColumn)
"""
expressions = []
conn = self._getConnection()
if self.ops.get('distinct'):
distinct = 'DISTINCT '
else:
distinct = ''
for func_name, attribute in attributes:
if not isinstance(attribute, str):
attribute = conn.sqlrepr(attribute)
expression = '%s(%s%s)' % (func_name, distinct, attribute)
expressions.append(expression)
return self.accumulate(*expressions)
def accumulateOne(self, func_name, attribute):
""" Making the sum/min/max/avg of a given select result attribute.
`attribute` can be a column name (like 'a_column')
or a dot-q attribute (like Table.q.aColumn)
"""
return self.accumulateMany((func_name, attribute))
def sum(self, attribute):
return self.accumulateOne("SUM", attribute)
def min(self, attribute):
return self.accumulateOne("MIN", attribute)
def avg(self, attribute):
return self.accumulateOne("AVG", attribute)
def max(self, attribute):
return self.accumulateOne("MAX", attribute)
def getOne(self, default=sqlbuilder.NoDefault):
"""
If a query is expected to only return a single value,
using ``.getOne()`` will return just that value.
If not results are found, ``SQLObjectNotFound`` will be
raised, unless you pass in a default value (like
``.getOne(None)``).
If more than one result is returned,
``SQLObjectIntegrityError`` will be raised.
"""
results = list(self)
if not results:
if default is sqlbuilder.NoDefault:
raise main.SQLObjectNotFound(
"No results matched the query for %s"
% self.sourceClass.__name__)
return default
if len(results) > 1:
raise main.SQLObjectIntegrityError(
"More than one result returned from query: %s"
% results)
return results[0]
def throughTo(self):
class _throughTo_getter(object):
def __init__(self, inst):
self.sresult = inst
def __getattr__(self, attr):
return self.sresult._throughTo(attr)
return _throughTo_getter(self)
throughTo = property(throughTo)
def _throughTo(self, attr):
otherClass = None
orderBy = sqlbuilder.NoDefault
ref = self.sourceClass.sqlmeta.columns.get(attr.endswith('ID') and attr or attr+'ID', None)
if ref and ref.foreignKey:
otherClass, clause = self._throughToFK(ref)
else:
join = [x for x in self.sourceClass.sqlmeta.joins if x.joinMethodName==attr]
if join:
join = join[0]
orderBy = join.orderBy
if hasattr(join, 'otherColumn'):
otherClass, clause = self._throughToRelatedJoin(join)
else:
otherClass, clause = self._throughToMultipleJoin(join)
if not otherClass:
raise AttributeError("throughTo argument (got %s) should be name of foreignKey or SQL*Join in %s" % (attr, self.sourceClass))
return otherClass.select(clause,
orderBy=orderBy,
connection=self._getConnection())
def _throughToFK(self, col):
otherClass = getattr(self.sourceClass, "_SO_class_"+col.foreignKey)
colName = col.name
query = self.queryForSelect().newItems([sqlbuilder.ColumnAS(getattr(self.sourceClass.q, colName), colName)]).orderBy(None).distinct()
query = sqlbuilder.Alias(query, "%s_%s" % (self.sourceClass.__name__, col.name))
return otherClass, otherClass.q.id==getattr(query.q, colName)
def _throughToMultipleJoin(self, join):
otherClass = join.otherClass
colName = join.soClass.sqlmeta.style.dbColumnToPythonAttr(join.joinColumn)
query = self.queryForSelect().newItems([sqlbuilder.ColumnAS(self.sourceClass.q.id, 'id')]).orderBy(None).distinct()
query = sqlbuilder.Alias(query, "%s_%s" % (self.sourceClass.__name__, join.joinMethodName))
joinColumn = getattr(otherClass.q, colName)
return otherClass, joinColumn==query.q.id
def _throughToRelatedJoin(self, join):
otherClass = join.otherClass
intTable = sqlbuilder.Table(join.intermediateTable)
colName = join.joinColumn
query = self.queryForSelect().newItems([sqlbuilder.ColumnAS(self.sourceClass.q.id, 'id')]).orderBy(None).distinct()
query = sqlbuilder.Alias(query, "%s_%s" % (self.sourceClass.__name__, join.joinMethodName))
clause = sqlbuilder.AND(otherClass.q.id == getattr(intTable, join.otherColumn),
getattr(intTable, colName) == query.q.id)
return otherClass, clause
|