gistfile1.diff

diff --git a/socorro/processor/processor.py b/socorro/processor/processor.py
index b3964a1..1754215 100755
--- a/socorro/processor/processor.py
+++ b/socorro/processor/processor.py
@@ -123,6 +123,8 @@ class Processor(object):
     self.prefixSignatureRegEx = re.compile(self.config.prefixSignatureRegEx)
     self.signaturesWithLineNumbersRegEx = re.compile(self.config.signaturesWithLineNumbersRegEx)
 
+    self.productIdMap = []
+
     self.reportsTable = sch.ReportsTable(logger=config.logger)
     self.extensionsTable = sch.ExtensionsTable(logger=config.logger)
     self.framesTable = sch.FramesTable(logger=config.logger)
@@ -170,7 +172,7 @@ class Processor(object):
       logger.debug("success")
       db_conn.commit()
     except sdb.exceptions_eligible_for_retry:
-      logger.debug("timout trouble")
+      logger.debug("timeout trouble")
       raise
     except sdb.db_module.ProgrammingError, x:
       logger.debug('the priority jobs table (%s) already exists',
@@ -656,6 +658,19 @@ class Processor(object):
     """
     #logger.debug("starting insertReportIntoDatabase")
     product = Processor.getJsonOrWarn(jsonDocument,'ProductName',processorErrorMessages,None, 30)
+    productId = Processor.getJsonOrWarn(jsonDocument,'ProductId',processorErrorMessages,None, 30)
+
+    # in some cases, we will receive products with the same name but different ID. cache the result.
+    if (not self.productIdMap):
+      self.productIdMap = self.loadProductIdMap()
+    logger.debug("product ID map: " + str(self.productIdMap))
+    logger.debug([x[0] for x in productIdMap])
+    for p in productIdMap:
+      if productId == p.productId:
+        if product != p.product and p.rewrite:
+          logger.info("Product name %s rewritten to %s", (product, p.product_name))
+          product = p.product_name
+
     version = Processor.getJsonOrWarn(jsonDocument,'Version', processorErrorMessages,None,16)
     buildID =   Processor.getJsonOrWarn(jsonDocument,'BuildID', processorErrorMessages,None,16)
     url = sutil.lookupLimitedStringOrNone(jsonDocument, 'URL', 255)
@@ -836,3 +851,27 @@ class Processor(object):
                                            showTraceback=False)
     except KeyError:
       self.config.logger.info('no Elastic Search URL has been configured')
+
+
+  #-----------------------------------------------------------------------------------------------------------------
+  def loadProductIdMap(self):
+    logger.debug('test1')
+    logger.debug(self.databaseConnectionPool.connectionCursorPair)
+    db_conn, db_cur = self.databaseConnectionPool.connectionCursorPair()
+    productIdMap = []
+    try:
+      logger.debug("attempting to load product_productid_map")
+      db_cur.execute("SELECT * FROM product_productid_map")
+      productIdList = db_cur.fetchall()
+      logger.debug("done loading product_productid_map")
+      db_conn.commit()
+    except:
+      logger.error('Unable to load product_productid_map')
+      db_conn.rollback()
+      raise
+    
+    columns = ('product_name', 'productid', 'rewrite', 'version_began', 'version_ended')
+    for result in productIdList:
+      productIdMap.append(dict(x for x in zip(columns, result)))
+
+    return productIdMap

30: Why the bracket?

31: Why assigning it to the instance? I.e. why not productIdMap = self.loadProductIdMap()?

50: The word "load" doesn't imply what it actually does. If it was prefixed "set" or "get" or "apply" I would understand what it does. At least, if it's called "load" then it shouldn't return anything.

51: What good does this do?

66: If you know the columns, stick that into the SELECT ... FROM ...

54: If it's called, a "somethingsomethingMap" I'd expect to be a dict type. Not a list.

61: When using commit/rollback I would much prefer a "try: finally:" approach simply because it's less about the error.

62: I'd prefer logger.error('Unable to load product_productid_map', exc_info=True) so you get the traceback in the logs.

35: Is p a dict instance coming from line 68? The p.productId would fail. However, p['productId'] would fail too because it's zipped up as p['productid']

57: In fact, if columns ever change, this script will fail because of the hardcoded columns tuple. Using the tuple instead of * would at least make us safe from newly added columns or change of internal order.