Improved exceptions

rowpipe/_meta.py

@@ -1,2 +1,2 @@
-__version__ = '0.1.0'
+__version__ = '0.1.1'
 __author__ = 'eric@civicknowledge.com'

rowpipe/exceptions.py

@@ -48,8 +48,14 @@ def __init__(self, type_target, field_header, value, message, *args, **kwargs):
         # Call the base class constructor with the parameters it needs
         Exception.__init__(self, textwrap.fill(message, 120), *args, **kwargs)
 
+
 class TooManyCastingErrors(RowPipeError):
-    pass
+
+    def __init__(self, *args, errors = None, **kwargs):
+        self.errors = errors if errors is not None else {}
+        Exception.__init__(self, *args, **kwargs)
+
+
 
 class SchemaError(RowPipeError):
     pass

rowpipe/table.py

@@ -48,9 +48,13 @@ def __init__(self, name, datatype=None, valuetype=None, transform=None):
             self.valuetype = resolve_value_type(valuetype)
         elif datatype is not None:
             self.valuetype = resolve_value_type(datatype)
+        else:
+            self.valuetype = None
 
         if self.valuetype is None:
-            raise SchemaError("Could not resovle type for datatype='{}' and valuetype='{}' ".format(datatype, valuetype))
+            raise SchemaError("Could not resovle type for for column '{}' datatype='{}' and valuetype='{}' "
+                              .format(name, datatype, valuetype))
+
 
 
         self.datatype = self.valuetype.python_type()

rowpipe/valuetype/core.py

@@ -48,8 +48,12 @@ def valuetype(func, *args, **kw):
 
 
 def count_errors(errors):
-    if sum(len(e) for e in errors.values()) > 50:
-        raise TooManyCastingErrors("Too many casting errors")
+
+    # Count at most 10 errors for each column
+    c = sum(len(e) if len(e) < 10 else 10 for e in errors.values())
+
+    if c > 50:
+        raise TooManyCastingErrors("Too many casting errors ({})".format(c), errors=errors)
 
 
 class TimeMixin(object):
@@ -203,7 +207,9 @@ def __getattr__(self, item):
 
 
 def cast_int(v, header_d, errors):
+    from rowpipe.valuetype import IntMeasure
     if isinstance(v, FailedValue):
+
         errors[header_d].add(u"Failed to cast '{}' ({}) to int in '{}': {}".format(v, type(v), header_d, v.exc))
         count_errors(errors)
         return None
@@ -522,7 +528,10 @@ def robust_int(v):
     if not v:
         return None
 
-    return int(v)
+    try:
+        return int(v)
+    except ValueError:
+        return int(float(v))
 
 
 def print_value(row_n, header_d, v):