$regex operator removed and $glob docu added

Author: Johannes Otepka

docs/source/reference/db.md

@@ -79,16 +79,22 @@ TaskRecord keys:
 
 MongoDB operators we emulate on all backends:
 
-| Operator | Python equivalent |
-| -------- | ----------------- |
-| '\$in'   | in                |
-| '\$nin'  | not in            |
-| '\$eq'   | ==                |
-| '\$ne'   | !=                |
-| '\$gt'   | >                 |
-| '\$gte'  | >=                |
-| '\$le'   | \<                |
-| '\$lte'  | \<=               |
+| Operator | Python equivalent                                                             |
+| -------- |-------------------------------------------------------------------------------|
+| '\$in'   | in                                                                            |
+| '\$nin'  | not in                                                                        |
+| '\$eq'   | ==                                                                            |
+| '\$ne'   | !=                                                                            |
+| '\$gt'   | >                                                                             |
+| '\$gte'  | >=                                                                            |
+| '\$le'   | \<                                                                            |
+| '\$lte'  | \<=                                                                           |
+| '\$glob' | [fnmatch](https://docs.python.org/3/library/fnmatch.html) (wildcard matching) | 
+
+Remarks on _$glob_: The operator can be used to find substrings in DB columns based on 
+[unix style filename pattern matching](https://docs.python.org/3/library/fnmatch.html) 
+_$glob_ is **not** a regular MongoDB opertor, but is internally translated to a regular 
+expression (_$regex_) which is natively supported by MongoDB.  
 
 The DB Query is useful for two primary cases:
 

ipyparallel/controller/dictdb.py

@@ -62,7 +62,6 @@
     '$mod': lambda a, b: a % b[0] == b[1],
     '$exists': lambda a, b: (b and a is not None) or (a is None and not b),
     '$glob': lambda a, b: fnmatch.fnmatch(a, b),
-    '$regex': lambda a, b: re.match(b, a),
 }
 
 

ipyparallel/controller/mongodb.py

@@ -125,7 +125,9 @@ def _translate(self, filter):
                     raise Exception(
                         f"unkown paramters {params.keys()} for %glob operator"
                     )
-                ret["$regex"] = fnmatch.translate(glob)
+                # we need to attach the start and end match character to achieve
+                # an equivalent matching behavior to fnmatch in mongoDB
+                ret["$regex"] = "^"+fnmatch.translate(glob)+"$"
             else:
                 for key, value in filter.items():
                     if isinstance(value, dict):

ipyparallel/controller/sqlitedb.py

@@ -49,7 +49,6 @@
     # '$mod': None,
     # '$exists' : None
     '$glob': "GLOB",
-    '$regex': "REGEXP",
 }
 null_operators = {
     '=': "IS NULL",

ipyparallel/tests/test_db.py

@@ -149,22 +149,6 @@ def test_find_records_glob(self):
         found = [r['msg_id'] for r in recs]
         assert set(ref) == set(found)
 
-    def test_find_records_regex(self):
-        """test finding records with '$regex' operators"""
-        hist = self.db.get_history()
-
-        pattern = "^.+_.$"
-        ref = [msg_id for msg_id in hist if re.match(pattern, msg_id)]
-        recs = self.db.find_records({'msg_id': {'$regex': pattern}}, ["msg_id"])
-        found = [r['msg_id'] for r in recs]
-        assert set(ref) == set(found)
-
-        pattern = "^.+_1[0-9]$"
-        ref = [msg_id for msg_id in hist if re.match(pattern, msg_id)]
-        recs = self.db.find_records({'msg_id': {'$regex': pattern}}, ["msg_id"])
-        found = [r['msg_id'] for r in recs]
-        assert set(ref) == set(found)
-
     def test_get_history(self):
         msg_ids = self.db.get_history()
         latest = datetime(1984, 1, 1).replace(tzinfo=utc)