relational/relational/optimizations.py
LtWorf 466c72ab2c - Other fixes
git-svn-id: http://galileo.dmi.unict.it/svn/relational/trunk@331 014f5005-505e-4b48-8d0a-63407b615a7c
2011-10-15 22:12:06 +00:00

590 lines
19 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# -*- coding: utf-8 -*-
# Relational
# Copyright (C) 2009 Salvo "LtWorf" Tomaselli
#
# Relational is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
# author Salvo "LtWorf" Tomaselli <tiposchi@tiscali.it>
#
# This module contains functions to perform various optimizations on the expression trees.
# The list general_optimizations contains pointers to general functions, so they can be called
# within a cycle.
#
# It is possible to add new general optimizations by adding the function in the list
# general_optimizations present in this module. And the optimization will be executed with the
# other ones when optimizing.
#
# A function will have one parameter, which is the root node of the tree describing the expression.
# The class used is defined in optimizer module.
# A function will have to return the number of changes performed on the tree.
import parser
from cStringIO import StringIO
from tokenize import generate_tokens
sel_op=('//=','**=','and','not','in','//','**','<<','>>','==','!=','>=','<=','+=','-=','*=','/=','%=','or','+','-','*','/','&','|','^','~','<','>','%','=','(',')',',','[',']')
PRODUCT=parser.PRODUCT
DIFFERENCE=parser.DIFFERENCE
UNION=parser.UNION
INTERSECTION=parser.INTERSECTION
DIVISION=parser.DIVISION
JOIN=parser.JOIN
JOIN_LEFT=parser.JOIN_LEFT
JOIN_RIGHT=parser.JOIN_RIGHT
JOIN_FULL=parser.JOIN_FULL
PROJECTION=parser.PROJECTION
SELECTION=parser.SELECTION
RENAME=parser.RENAME
ARROW=parser.ARROW
def replace_node(replace,replacement):
'''This function replaces "replace" node with the node "with",
the father of the node will now point to the with node'''
replace.name=replacement.name
replace.kind=replacement.kind
if replace.kind==parser.UNARY:
replace.child=replacement.child
replace.prop=replacement.prop
elif replace.kind==parser.BINARY:
replace.right=replacement.right
replace.left=replacement.left
def recoursive_scan(function,node,rels=None):
'''Does a recoursive optimization on the tree.
This function will recoursively execute the function given
as "function" parameter starting from node to all the tree.
if rels is provided it will be passed as argument to the function.
Otherwise the function will be called just on the node.
Result value: function is supposed to return the amount of changes
it has performed on the tree.
The various result will be added up and this final value will be the
returned value.'''
changes=0
#recoursive scan
if node.kind==parser.UNARY:
if rels!=None:
changes+=function(node.child,rels)
else:
changes+=function(node.child)
elif node.kind==parser.BINARY:
if rels!=None:
changes+=function(node.right,rels)
changes+=function(node.left,rels)
else:
changes+=function(node.right)
changes+=function(node.left)
return changes
def duplicated_select(n):
'''This function locates and deletes things like
σ a ( σ a(C)) and the ones like σ a ( σ b(C))
replacing the 1st one with a single select and
the 2nd one with a single select with both conditions
in and
'''
changes=0
if n.name==SELECTION and n.child.name==SELECTION:
if n.prop != n.child.prop: #Nested but different, joining them
n.prop = n.prop + " and " + n.child.prop
#This adds parenthesis if they are needed
if n.child.prop.startswith('(') or n.prop.startswith('('):
n.prop= '(%s)' % n.prop
n.child=n.child.child
changes=1
changes+=duplicated_select(n)
return changes+recoursive_scan(duplicated_select,n)
def futile_union_intersection_subtraction(n):
'''This function locates things like r r, and replaces them with r.
R R --> R
R ᑎ R --> R
R - R --> σ False (R)
σ k (R) - R --> σ False (R)
R - σ k (R) --> σ not k (R)
σ k (R) R --> R
σ k (R) ᑎ R --> σ k (R)
'''
changes=0
#Union and intersection of the same thing
if n.name in (UNION,INTERSECTION) and n.left==n.right:
changes=1
replace_node(n,n.left)
#selection and union of the same thing
elif (n.name == UNION):
if n.left.name==SELECTION and n.left.child==n.right:
changes=1
replace_node(n,n.right)
elif n.right.name==SELECTION and n.right.child==n.left:
changes=1
replace_node(n,n.left)
#selection and intersection of the same thing
elif n.name == INTERSECTION:
if n.left.name==SELECTION and n.left.child==n.right:
changes=1
replace_node(n,n.left)
elif n.right.name==SELECTION and n.right.child==n.left:
changes=1
replace_node(n,n.right)
#Subtraction and selection of the same thing
elif (n.name == DIFFERENCE and (n.right.name==SELECTION and n.right.child==n.left)): #Subtraction of two equal things, but one has a selection
n.name=n.right.name
n.kind=n.right.kind
n.child=n.right.child
n.prop='(not (%s))' % n.right.prop
n.left=n.right=None
#Subtraction of the same thing or with selection on the left child
elif (n.name==DIFFERENCE and ((n.left==n.right) or (n.left.name==SELECTION and n.left.child==n.right)) ):#Empty relation
changes=1
n.kind=parser.UNARY
n.name=SELECTION
n.prop='False'
n.child=n.left.get_left_leaf()
#n.left=n.right=None
return changes+recoursive_scan(futile_union_intersection_subtraction,n)
def down_to_unions_subtractions_intersections(n):
'''This funcion locates things like σ i==2 (c d), where the union
can be a subtraction and an intersection and replaces them with
σ i==2 (c) σ i==2(d).
'''
changes=0
_o=(UNION,DIFFERENCE,INTERSECTION)
if n.name==SELECTION and n.child.name in _o:
left=parser.node()
left.prop=n.prop
left.name=n.name
left.child=n.child.left
left.kind=parser.UNARY
right=parser.node()
right.prop=n.prop
right.name=n.name
right.child=n.child.right
right.kind=parser.UNARY
n.name=n.child.name
n.left=left
n.right=right
n.child=None
n.prop=None
n.kind=parser.BINARY
changes+=1
return changes+recoursive_scan(down_to_unions_subtractions_intersections,n)
def duplicated_projection(n):
'''This function locates thing like π i ( π j (R)) and replaces
them with π i (R)'''
changes=0
if n.name==PROJECTION and n.child.name==PROJECTION:
n.child=n.child.child
changes+=1
return changes+recoursive_scan(duplicated_projection,n)
def selection_inside_projection(n):
'''This function locates things like σ j (π k(R)) and
converts them into π k(σ j (R))'''
changes=0
if n.name==SELECTION and n.child.name==PROJECTION:
changes=1
temp=n.prop
n.prop=n.child.prop
n.child.prop=temp
n.name=PROJECTION
n.child.name=SELECTION
return changes+recoursive_scan(selection_inside_projection,n)
def swap_union_renames(n):
'''This function locates things like
ρ a➡b(R) ρ a➡b(Q)
and replaces them with
ρ a➡b(R Q).
Does the same with subtraction and intersection'''
changes=0
if n.name in (DIFFERENCE,UNION,INTERSECTION) and n.left.name==n.right.name and n.left.name==RENAME:
l_vars={}
for i in n.left.prop.split(','):
q=i.split(ARROW)
l_vars[q[0].strip()]=q[1].strip()
r_vars={}
for i in n.right.prop.split(','):
q=i.split(ARROW)
r_vars[q[0].strip()]=q[1].strip()
if r_vars==l_vars:
changes=1
#Copying self, but child will be child of renames
q=parser.node()
q.name=n.name
q.kind=parser.BINARY
q.left=n.left.child
q.right=n.right.child
n.name=RENAME
n.kind=parser.UNARY
n.child=q
n.prop=n.left.prop
n.left=n.right=None
return changes+recoursive_scan(swap_union_renames,n)
def futile_renames(n):
'''This function purges renames like id->id'''
changes=0
if n.name==RENAME:
#Located two nested renames.
changes=1
#Creating a dictionary with the attributes
_vars={}
for i in n.prop.split(','):
q=i.split(ARROW)
_vars[q[0].strip()]=q[1].strip()
#Scans dictionary to locate things like "a->b,b->c" and replace them with "a->c"
for key in list(_vars.keys()):
try:
value=_vars[key]
except:
value=None
if key==value:
_vars.pop(value) #Removes the unused one
#Reset prop var
n.prop=""
#Generates new prop var
for i in _vars.items():
n.prop+=u"%s%s%s," % (i[0],ARROW,i[1])
n.prop=n.prop[:-1] #Removing ending comma
if len(n.prop)==0: #Nothing to rename, removing the rename op
replace_node(n,n.child)
return changes+recoursive_scan(futile_renames,n)
def subsequent_renames(n):
'''This function removes redoundant subsequent renames joining them into one'''
'''Purges renames like id->id Since it's needed to be performed BEFORE this one
so it is not in the list with the other optimizations'''
futile_renames(n)
changes=0
if n.name==RENAME and n.child.name==n.name:
#Located two nested renames.
changes=1
#Joining the attribute into one
n.prop+=','+n.child.prop
n.child=n.child.child
#Creating a dictionary with the attributes
_vars={}
for i in n.prop.split(','):
q=i.split(ARROW)
_vars[q[0].strip()]=q[1].strip()
#Scans dictionary to locate things like "a->b,b->c" and replace them with "a->c"
for key in list(_vars.keys()):
try:
value=_vars[key]
except:
value=None
if value in _vars.keys():
if _vars[value]!=key:
#Double rename on attribute
_vars[key] = _vars[_vars[key]] #Sets value
_vars.pop(value) #Removes the unused one
else: #Cycle rename a->b,b->a
_vars.pop(value) #Removes the unused one
_vars.pop(key) #Removes the unused one
#Reset prop var
n.prop=""
#Generates new prop var
for i in _vars.items():
n.prop+=u"%s%s%s," % (i[0],ARROW,i[1])
n.prop=n.prop[:-1] #Removing ending comma
if len(n.prop)==0: #Nothing to rename, removing the rename op
replace_node(n,n.child)
return changes+recoursive_scan(subsequent_renames,n)
class level_string(str):
level=0
def tokenize_select(expression):
'''This function returns the list of tokens present in a
selection. The expression can contain parenthesis.
It will use a subclass of str with the attribute level, which
will specify the nesting level of the token into parenthesis.'''
g=generate_tokens(StringIO(str(expression)).readline)
l=list(token[1] for token in g)
l.remove('')
#Changes the 'a','.','method' token group into a single 'a.method' token
try:
while True:
dot=l.index('.')
l[dot]='%s.%s' % (l[dot-1],l[dot+1])
l.pop(dot+1)
l.pop(dot-1)
except:
pass
level=0
for i in range(len(l)):
l[i]=level_string(l[i])
l[i].level=level
if l[i]=='(':
level+=1
elif l[i]==')':
level-=1
return l
def swap_rename_projection(n):
'''This function locates things like π k(ρ j(R))
and replaces them with ρ j(π k(R)).
This will let rename work on a hopefully smaller set
and more important, will hopefully allow further optimizations.
Will also eliminate fields in the rename that are cutted in the projection.
'''
changes=0
if n.name==PROJECTION and n.child.name==RENAME:
changes=1
#π index,name(ρ id➡index(R))
_vars={}
for i in n.child.prop.split(','):
q=i.split(ARROW)
_vars[q[1].strip()]=q[0].strip()
_pr=n.prop.split(',')
for i in range(len(_pr)):
try:
_pr[i]=_vars[_pr[i].strip()]
except:
pass
_pr_reborn=n.prop.split(',')
for i in list(_vars.keys()):
if i not in _pr_reborn:
_vars.pop(i)
n.name=n.child.name
n.prop=''
for i in _vars.keys():
n.prop+=u'%s%s%s,' % (_vars[i],ARROW,i)
n.prop=n.prop[:-1]
n.child.name=PROJECTION
n.child.prop=''
for i in _pr:
n.child.prop+=i+','
n.child.prop=n.child.prop[:-1]
return changes+recoursive_scan(swap_rename_projection,n)
def swap_rename_select(n):
'''This function locates things like σ k(ρ j(R)) and replaces
them with ρ j(σ k(R)). Renaming the attributes used in the
selection, so the operation is still valid.'''
changes=0
if n.name==SELECTION and n.child.name==RENAME:
changes=1
#Dictionary containing attributes of rename
_vars={}
for i in n.child.prop.split(','):
q=i.split(ARROW)
_vars[q[1].strip()]=q[0].strip()
#tokenizes expression in select
_tokens=tokenize_select(n.prop)
#Renaming stuff
for i in range(len(_tokens)):
splitted=_tokens[i].split('.',1)
if splitted[0] in _vars:
if len(splitted)==1:
_tokens[i]=_vars[_tokens[i].split('.')[0]]
else:
_tokens[i]=_vars[_tokens[i].split('.')[0]]+'.'+splitted[1]
#Swapping operators
n.name=RENAME
n.child.name=SELECTION
n.prop=n.child.prop
n.child.prop=''
for i in _tokens:
n.child.prop+=i+ ' '
return changes+recoursive_scan(swap_rename_select,n)
def select_union_intersect_subtract(n):
'''This function locates things like σ i(a) σ q(a)
and replaces them with σ (i OR q) (a)
Removing a O(n²) operation like the union'''
changes=0
if n.name in (UNION, INTERSECTION, DIFFERENCE) and n.left.name==SELECTION and n.right.name==SELECTION and n.left.child==n.right.child:
cahnges=1
d={UNION:'or', INTERSECTION:'and', DIFFERENCE:'and not'}
op=d[n.name]
newnode=parser.node()
newnode.prop='((%s) %s (%s))' % (n.left.prop,op,n.right.prop)
newnode.name=SELECTION
newnode.child=n.left.child
newnode.kind=parser.UNARY
replace_node(n,newnode)
return changes+recoursive_scan(select_union_intersect_subtract,n)
def selection_and_product(n,rels):
'''This function locates things like σ k (R*Q) and converts them into
σ l (σ j (R) * σ i (Q)). Where j contains only attributes belonging to R,
i contains attributes belonging to Q and l contains attributes belonging to both'''
changes=0
if n.name==SELECTION and n.child.name in (PRODUCT,JOIN,JOIN_LEFT,JOIN_RIGHT,JOIN_FULL):
l_attr=n.child.left.result_format(rels)
r_attr=n.child.right.result_format(rels)
tokens=tokenize_select(n.prop)
groups=[]
temp=[]
for i in tokens:
if i=='and' and i.level==0:
groups.append(temp)
temp=[]
else:
temp.append(i)
if len(temp)!=0:
groups.append(temp)
temp=[]
left=[]
right=[]
both=[]
for i in groups:
l_fields=False #has fields in left?
r_fields=False #has fields in left?
for j in set(i).difference(sel_op):
j=j.split('.')[0]
if j in l_attr:#Field in left
l_fields=True
if j in r_attr:#Field in right
r_fields=True
if l_fields and r_fields:#Fields in both
both.append(i)
elif l_fields:
left.append(i)
elif r_fields:
right.append(i)
else:#Unknown.. adding in both
both.append(i)
#Preparing left selection
if len(left)>0:
changes=1
l_node=parser.node()
l_node.name=SELECTION
l_node.kind=parser.UNARY
l_node.child=n.child.left
l_node.prop=''
n.child.left=l_node
while len(left)>0:
c=left.pop(0)
for i in c:
l_node.prop+=i+ ' '
if len(left)>0:
l_node.prop+=' and '
if '(' in l_node.prop:
l_node.prop='(%s)' % l_node.prop
#Preparing right selection
if len(right)>0:
changes=1
r_node=parser.node()
r_node.name=SELECTION
r_node.prop=''
r_node.kind=parser.UNARY
r_node.child=n.child.right
n.child.right=r_node
while len(right)>0:
c=right.pop(0)
for i in c:
r_node.prop+=i+ ' '
if len(right)>0:
r_node.prop+=' and '
if '(' in r_node.prop:
r_node.prop='(%s)' % r_node.prop
#Changing main selection
n.prop=''
if len(both)!=0:
while len(both)>0:
c=both.pop(0)
for i in c:
n.prop+=i+ ' '
if len(both)>0:
n.prop+=' and '
if '(' in n.prop:
n.prop='(%s)' % n.prop
else:#No need for general select
replace_node(n,n.child)
return changes+recoursive_scan(selection_and_product,n,rels)
general_optimizations=[duplicated_select,down_to_unions_subtractions_intersections,duplicated_projection,selection_inside_projection,subsequent_renames,swap_rename_select,futile_union_intersection_subtraction,swap_union_renames,swap_rename_projection,select_union_intersect_subtract]
specific_optimizations=[selection_and_product]
if __name__=="__main__":
print tokenize_select("skill == 'C' and id % 2 == 0")