I have a simple data fow task that consists of a Flat File Connection
linked to a script component, before moving to a Data Flow Destination.

The following are true:

I have setup the Flat File Source with the metadata which happens to be
all string [DT_STR] with different OutputColumnWidth values.

I have set Error Output for ALL errors for the Flat File Connection to
redirect rows to another Flat File Connection.

I set the MaximumErrorCount for the Data Flow Task to 1000.

My script component performs some simple string validation of the data
before it sends it to the Data Flow Destination.



When I run the package it imports most of the rows fine BUT then there
is that dirty piece of data that causes alot of headaches that makes
the script component fail with the following error message: "The value
is too large to fit in the column data area of the buffer."

Questions:

1. If the value is too large (it has to be a string that is to long and
therefore exceedes the set size for that column) how come it got
imported by the Flat File Connection and not redirected as a truncation
error?

2. Why did the package not continue since I set MaximumErrorCount to
1000?


I obviously DO NOT want to pass data along to the script component that
is "too large" and cause everythign to fail. Isn't that what the
metadata is for? Isn't that what the Error RedirectRow is supposed to
take care of? I am so frustrated, can someone please let me know how I
can avoid this scenario going forward.

Thanks