debug

provsql/src/provenance.py

@@ -14,14 +14,37 @@ import traceback
 tablesWithProvenance = {}
 
 def tablesInQuery(query):
-    parsed = list(sp.parse(query))
-    tables = []
-    for s in parsed:
-        if s.get_type() != 'UNKNOWN':
-            output_t = ut.get_table_identifier(s)
-            tables.append(set(ut.decompose_identifiers(output_t)))
-
-    return tables
+    parsed = sp.parse(query)[0]
+    output = []
+    state = ""
+    for token in parsed.tokens:
+        cur = token.value
+        if cur == " ":
+            continue
+        if cur.lower() == "from":
+            state = "first table"
+            continue
+
+        if state == "first table":
+            if "(" not in cur and "join" in cur:
+                state = "second table"
+                continue
+            if "(" in cur and ")" in cur:
+                #subquery
+                output.extend(tablesInQuery(cur[cur.index("(")+1 : len(cur) - 1 - cur[::-1].index(")")]))
+            else:
+                output.append(cur.split("as")[0].strip())
+            continue
+        
+        if state == "second table":
+            if "(" in cur and ")" in cur:
+                #subquery
+                output.extend(tablesInQuery(cur[cur.index("(")+1 : len(cur) - 1 - cur[::-1].index(")")]))
+            else:
+                output.append(cur.split("as")[0].strip())
+            break
+
+    return list(set(output))
 
 
 #Preprocess query

provsql/src/provenanceUtils.py

-# Code for table processing is from Haorong (haorong4@illinois.edu). Thanks to Haorong.
-#  
-
 def commitAndExecute(conn, cur, command):
     cur.execute(command)
     conn.commit()
@@ -94,55 +91,4 @@ def combineSameRow(input):
             output.append(row)
             lastseen = row
     return output
-        
-# Following code is from Haorong
-############################################################################
-
-from sqlparse.sql import IdentifierList, Identifier,Parenthesis, Token, Comparison, Where, Function
-from sqlparse.tokens import Keyword, DML,Whitespace, Punctuation, Name
-
-def is_subquery(parsed):
-    if not parsed.is_group:
-        return False
-    for item in parsed.tokens:
-        if item.ttype is DML and item.value.upper() == 'SELECT':
-            return True
-    return False
-
-def get_table_identifier(parsed):
-    from_flag = False
-    view_flag = False
-    result = []
-    for item in list(parsed.tokens):
-        if item.is_group:
-            for x in get_table_identifier(item):
-                result.append(x)
-        if from_flag:
-            if is_subquery(item):
-                for x in get_table_identifier(item):
-                    result.append(x)
-            elif item.ttype is Keyword and item.value.upper() in ['ORDER', 'GROUP', 'BY', 'HAVING', 'GROUP BY', 'ON', 'WHERE', 'SET']:
-                from_flag = False
-                break
-            else:
-                result.append(item)
-        elif view_flag and not item.ttype == Whitespace:
-            result.append(item)
-            view_flag = False
-        
-        if (item.ttype is Keyword and item.value.upper()in ['FROM', 'INTO'] ) or ( item.ttype is DML and item.value.upper() == 'UPDATE' ):
-            from_flag = True
-        if item.ttype is Keyword and item.value.upper() == 'VIEW':
-            view_flag = True
-    return result
-
-def decompose_identifiers(identifiers):
-    result = []
-    for item in identifiers:
-        if isinstance(item, IdentifierList):
-            for identifier in item.get_identifiers():               
-                result.append(item)
-        else:         
-            result.append(item)
-    return result