/[thuban]/branches/WIP-pyshapelib-bramz/Thuban/Model/transientdb.py
ViewVC logotype

Diff of /branches/WIP-pyshapelib-bramz/Thuban/Model/transientdb.py

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 777 by bh, Tue Apr 29 14:54:12 2003 UTC revision 849 by bh, Wed May 7 11:55:31 2003 UTC
# Line 85  class ColumnReference: Line 85  class ColumnReference:
85          self.internal_name = internal_name          self.internal_name = internal_name
86    
87    
88  class TransientTableBase:  class TransientTableBase(table.OldTableInterfaceMixin):
89    
90      """Base class for tables in the transient database"""      """Base class for tables in the transient database"""
91    
# Line 96  class TransientTableBase: Line 96  class TransientTableBase:
96          self.indexed_columns = {}          self.indexed_columns = {}
97          self.read_record_cursor = None          self.read_record_cursor = None
98          self.read_record_last_row = None          self.read_record_last_row = None
99            self.read_record_last_result = None
100    
101      def create(self, columns):      def create(self, columns):
102          self.columns = columns          self.columns = columns
# Line 103  class TransientTableBase: Line 104  class TransientTableBase:
104          self.orig_names = []          self.orig_names = []
105          self.internal_to_orig = {}          self.internal_to_orig = {}
106          self.orig_to_internal = {}          self.orig_to_internal = {}
107            self.column_map = {}
108    
109          # Create the column objects and fill various maps and lists          # Create the column objects and fill various maps and lists
110          for col in self.columns:          for index in range(len(self.columns)):
111                col = self.columns[index]
112              self.name_to_column[col.name] = col              self.name_to_column[col.name] = col
113              self.orig_names.append(col.name)              self.orig_names.append(col.name)
114              self.internal_to_orig[col.internal_name] = col.name              self.internal_to_orig[col.internal_name] = col.name
115              self.orig_to_internal[col.name] = col.internal_name              self.orig_to_internal[col.name] = col.internal_name
116                self.column_map[col.name] = col
117                self.column_map[index] = col
118    
119          # Build the CREATE TABLE statement and create the table in the          # Build the CREATE TABLE statement and create the table in the
120          # database          # database
121          table_types = []          table_types = ["id INTEGER PRIMARY KEY"]
122          for col in self.columns:          for col in self.columns:
123              table_types.append("%s %s" % (col.internal_name,              table_types.append("%s %s" % (col.internal_name,
124                                            sql_type_map[col.type]))                                            sql_type_map[col.type]))
# Line 137  class TransientTableBase: Line 142  class TransientTableBase:
142              self.db.execute(stmt)              self.db.execute(stmt)
143              self.indexed_columns[column] = 1              self.indexed_columns[column] = 1
144    
145      def field_count(self):      def NumColumns(self):
146          return len(self.columns)          return len(self.columns)
147    
148      def field_info(self, i):      def NumRows(self):
         col = self.columns[i]  
         return col.type, col.name, 0, 0  
   
     def field_info_by_name(self, name):  
         for col in self.columns:  
             if col.name == name:  
                 return col.type, col.name, 0, 0  
         else:  
             return None  
   
     def record_count(self):  
149          result = self.db.execute("SELECT count(*) FROM %s;" % self.tablename)          result = self.db.execute("SELECT count(*) FROM %s;" % self.tablename)
150          return int(result[0])          return int(result[0])
151    
152      def read_record(self, index):      def Columns(self):
153            return self.columns
154    
155        def Column(self, col):
156            return self.column_map[col]
157    
158        def HasColumn(self, col):
159            """Return whether the table has a column with the given name or index
160            """
161            return self.column_map.has_key(col)
162    
163        def ReadRowAsDict(self, index):
164            # Implementation Strategy: Executing a completely new select
165            # statement every time this method is called is too slow. The
166            # most important usage is to read the records more or less
167            # sequentially. This happens e.g. when drawing a layer with a
168            # classification where the shapes are drawn in order of the
169            # shape ids. Another pattern is that the same row is requested
170            # several times in a row. This happens in the table view, for
171            # instance.
172    
173            # We can exploit this to make access faster by having one cursor
174            # open all the time and keeping the last row read around in case
175            # the same row is accessed again the next time and if the row
176            # index is larger than the row we have read last we simply fetch
177            # rows from the cursor until we've reached the requested row. If
178            # the requested row index is smaller then we start a new cursor.
179    
180            # FIXME: So far this scheme seems to work well enough. Obvious
181            # improvements would be to start the cursor at exactly the
182            # requested row (should be efficient and easy to do now that the
183            # id is the primary key) and to perhaps to also start a new
184            # cursor if the requested index is much larger than the last row
185            # so that we don't read and discard lots of the rows.
186    
187            # Check whether we have to start a new cursor
188          if self.read_record_cursor is None or index <self.read_record_last_row:          if self.read_record_cursor is None or index <self.read_record_last_row:
189              stmt = ("SELECT %s FROM %s;"              stmt = ("SELECT %s FROM %s;"
190                      % (", ".join([c.internal_name for c in self.columns]),                      % (", ".join([c.internal_name for c in self.columns]),
# Line 163  class TransientTableBase: Line 192  class TransientTableBase:
192              self.read_record_cursor = self.db.cursor()              self.read_record_cursor = self.db.cursor()
193              self.read_record_cursor.execute(stmt)              self.read_record_cursor.execute(stmt)
194              self.read_record_last_row = -1              self.read_record_last_row = -1
195          for i in range(index - self.read_record_last_row):              self.read_record_last_result = None
196              result = self.read_record_cursor.fetchone()  
197            # Now we should have a cursor at a position less than or equal
198            # to the index so the following if statement will always set
199            # result to a suitable value
200            assert index >= self.read_record_last_row
201    
202            if index == self.read_record_last_row:
203                result = self.read_record_last_result
204            else:
205                for i in range(index - self.read_record_last_row):
206                    result = self.read_record_cursor.fetchone()
207                    self.read_record_last_result = result
208          self.read_record_last_row = index          self.read_record_last_row = index
209          result = dict(zip(self.orig_names, result))          return dict(zip(self.orig_names, result))
210          return result  
211        def ReadValue(self, row, col):
212            """Return the value of the specified row and column
213    
214            The col parameter may be the index of the column or its name.
215            """
216            # Depending on the actual access patterns of the table data, it
217            # might be a bit faster in some circumstances to not implement
218            # this via ReadRowAsDict, but this simple implementation should
219            # be fast enough for most purposes.
220            return self.ReadRowAsDict(row)[self.column_map[col].name]
221    
222      def field_range(self, colname):      def ValueRange(self, col):
223          col = self.name_to_column[colname]          col = self.column_map[col]
224          iname = col.internal_name          iname = col.internal_name
225          min, max = self.db.execute("SELECT min(%s), max(%s) FROM %s;"          min, max = self.db.execute("SELECT min(%s), max(%s) FROM %s;"
226                                     % (iname, iname, self.tablename))                                     % (iname, iname, self.tablename))
227          converter = type_converter_map[col.type]          converter = type_converter_map[col.type]
228          return ((converter(min), None), (converter(max), None))          return (converter(min), converter(max))
229    
230      def GetUniqueValues(self, colname):      def UniqueValues(self, col):
231          iname = self.orig_to_internal[colname]          iname = self.column_map[col].internal_name
232          cursor = self.db.cursor()          cursor = self.db.cursor()
233          cursor.execute("SELECT %s FROM %s GROUP BY %s;"          cursor.execute("SELECT %s FROM %s GROUP BY %s;"
234                         % (iname, self.tablename, iname))                         % (iname, self.tablename, iname))
# Line 190  class TransientTableBase: Line 240  class TransientTableBase:
240              result.append(row[0])              result.append(row[0])
241          return result          return result
242    
243        def SimpleQuery(self, left, comparison, right):
244            """Return the indices of all rows that matching a condition.
245    
246            Parameters:
247               left -- The column object for the left side of the comparison
248    
249               comparison -- The comparison operator as a string. It must be
250                             one of '==', '!=', '<', '<=', '>=', '>'
251    
252               right -- The right hand side of the comparison. It must be
253                        either a column object or a value, i.e. a string,
254                        int or float.
255    
256            The return value is a sorted list of the indices of the rows
257            where the condition is true.
258            """
259            if comparison not in ("==", "!=", "<", "<=", ">=", ">"):
260                raise ValueError("Comparison operator %r not allowed" % comparison)
261    
262            if hasattr(right, "internal_name"):
263                right_template = right.internal_name
264                params = ()
265            else:
266                right_template = "%s"
267                params = (right,)
268    
269            query = "SELECT id FROM %s WHERE %s %s %s ORDER BY id;" \
270                    % (self.tablename, left.internal_name, comparison,
271                       right_template)
272    
273            cursor = self.db.cursor()
274            cursor.execute(query, params)
275            result = []
276            while 1:
277                row = cursor.fetchone()
278                if row is None:
279                    break
280                result.append(row[0])
281            return result
282    
283    
284  class TransientTable(TransientTableBase):  class TransientTable(TransientTableBase):
285    
# Line 206  class TransientTable(TransientTableBase) Line 296  class TransientTable(TransientTableBase)
296    
297      def create(self, table):      def create(self, table):
298          columns = []          columns = []
299          for i in range(table.field_count()):          for col in table.Columns():
300              type, name = table.field_info(i)[:2]              columns.append(ColumnReference(col.name, col.type,
             columns.append(ColumnReference(name, type,  
301                                             self.db.new_column_name()))                                             self.db.new_column_name()))
302          TransientTableBase.create(self, columns)          TransientTableBase.create(self, columns)
303    
304          # copy the input table to the transient db          # copy the input table to the transient db
305          insert_template = "INSERT INTO %s (%s) VALUES (%s);" \  
306            # A key to insert to use for the formatting of the insert
307            # statement. The key must not be equal to any of the column
308            # names so we construct one by building a string of x's that is
309            # longer than any of the column names
310            id_key = max([len(col.name) for col in self.columns]) * "x"
311    
312            insert_template = "INSERT INTO %s (id, %s) VALUES (%%(%s)s, %s);" \
313                                 % (self.tablename,                                 % (self.tablename,
314                                    ", ".join([col.internal_name                                    ", ".join([col.internal_name
315                                               for col in self.columns]),                                               for col in self.columns]),
316                                      id_key,
317                                    ", ".join(["%%(%s)s" % col.name                                    ", ".join(["%%(%s)s" % col.name
318                                               for col in self.columns]))                                               for col in self.columns]))
319          cursor = self.db.cursor()          cursor = self.db.cursor()
320          for i in range(table.record_count()):          for i in range(table.NumRows()):
321              cursor.execute(insert_template, table.read_record(i))              row = table.ReadRowAsDict(i)
322                row[id_key] = i
323                cursor.execute(insert_template, row)
324          self.db.conn.commit()          self.db.conn.commit()
325    
326    
# Line 268  class TransientJoinedTable(TransientTabl Line 367  class TransientJoinedTable(TransientTabl
367          columns = []          columns = []
368          for col in self.left_table.columns + self.right_table.columns:          for col in self.left_table.columns + self.right_table.columns:
369              if col.name in visited:              if col.name in visited:
370                    # We can't allow multiple columns with the same original
371                    # name, so omit this one. FIXME: There should be a
372                    # better solution.
373                  continue                  continue
374              columns.append(col)              columns.append(col)
375          TransientTableBase.create(self, columns)          TransientTableBase.create(self, columns)
376    
377          # Copy the joined data to the table.          # Copy the joined data to the table.
378          internal_names = [col.internal_name for col in self.columns]          internal_names = [col.internal_name for col in self.columns]
379          stmt = "INSERT INTO %s (%s) SELECT %s FROM %s JOIN %s ON %s = %s;" \          stmt = ("INSERT INTO %s (id, %s) SELECT %s.id, %s FROM %s"
380                 % (self.tablename,                  " JOIN %s ON %s = %s;"
381                    ", ".join(internal_names),                  % (self.tablename,
382                    ", ".join(internal_names),                     ", ".join(internal_names),
383                    self.left_table.tablename,                     self.left_table.tablename,
384                    self.right_table.tablename,                     ", ".join(internal_names),
385                    self.orig_to_internal[self.left_field],                     self.left_table.tablename,
386                    self.orig_to_internal[self.right_field])                     self.right_table.tablename,
387                       self.orig_to_internal[self.left_field],
388                       self.orig_to_internal[self.right_field]))
389          self.db.execute(stmt)          self.db.execute(stmt)
390    
391    
392  class AutoTransientTable:  class AutoTransientTable(table.OldTableInterfaceMixin):
393    
394      """Table that copies data to a transient table on demand.      """Table that copies data to a transient table on demand.
395    
# Line 298  class AutoTransientTable: Line 402  class AutoTransientTable:
402          self.table = table          self.table = table
403          self.t_table = None          self.t_table = None
404    
405      def record_count(self):      def Columns(self):
406          """Return the number of records"""          return self.table.Columns()
         return self.table.record_count()  
   
     def field_count(self):  
         """Return the number of fields in a record"""  
         return self.table.field_count()  
   
     def field_info(self, field):  
         """Return a tuple (type, name, width, prec) for the field no. field  
   
         type is the data type of the field, name the name, width the  
         field width in characters and prec the decimal precision.  
         """  
         info = self.table.field_info(field)  
         if info:  
             info = info[:2] + (0, 0)  
         return info  
   
     def field_info_by_name(self, fieldName):  
         info = self.table.field_info_by_name(fieldName)  
         if info:  
             info = info[:2] + (0, 0)  
         return info  
407    
408      def read_record(self, record):      def Column(self, col):
409            return self.table.Column(col)
410    
411        def HasColumn(self, col):
412            """Return whether the table has a column with the given name or index
413            """
414            return self.table.HasColumn(col)
415    
416        def NumRows(self):
417            return self.table.NumRows()
418    
419        def NumColumns(self):
420            return self.table.NumColumns()
421    
422        def ReadRowAsDict(self, record):
423          """Return the record no. record as a dict mapping field names to values          """Return the record no. record as a dict mapping field names to values
424          """          """
425          if self.t_table is not None:          if self.t_table is not None:
426              return self.t_table.read_record(record)              return self.t_table.ReadRowAsDict(record)
427          else:          else:
428              return self.table.read_record(record)              return self.table.ReadRowAsDict(record)
429    
430        def ReadValue(self, row, col):
431            """Return the value of the specified row and column
432    
433      def write_record(self, record, values):          The col parameter may be the index of the column or its name.
434          raise NotImplementedError          """
435            if self.t_table is not None:
436                return self.t_table.ReadValue(row, col)
437            else:
438                return self.table.ReadValue(row, col)
439    
440      def copy_to_transient(self):      def copy_to_transient(self):
441          """Internal: Create a transient table and copy the data into it"""          """Internal: Create a transient table and copy the data into it"""
# Line 346  class AutoTransientTable: Line 449  class AutoTransientTable:
449              self.copy_to_transient()              self.copy_to_transient()
450          return self.t_table          return self.t_table
451    
452      def field_range(self, colname):      def ValueRange(self, col):
453            if self.t_table is None:
454                self.copy_to_transient()
455            return self.t_table.ValueRange(col)
456    
457        def UniqueValues(self, col):
458          if self.t_table is None:          if self.t_table is None:
459              self.copy_to_transient()              self.copy_to_transient()
460          return self.t_table.field_range(colname)          return self.t_table.UniqueValues(col)
461    
462      def GetUniqueValues(self, colname):      def SimpleQuery(self, left, comparison, right):
463          if self.t_table is None:          if self.t_table is None:
464              self.copy_to_transient()              self.copy_to_transient()
465          return self.t_table.GetUniqueValues(colname)          # Make sure to use the column object of the transient table. The
466            # left argument is always a column object so we can just ask the
467            # t_table for the right object.
468            return self.t_table.SimpleQuery(self.t_table.Column(left.name),
469                                            comparison, right)
Legend:
Removed from v.777  
changed lines
  Added in v.849
[email protected]
 ViewVC Help
Powered by ViewVC 1.1.26