Calling all Access geeks ...
Is there a legitimate way to explain the Access problem that has been described in the Wisconsin precinct that just "found" 14,315 missing votes that "failed to save" from the the city of Brookfield ... ???
Wisconsin stunner: Prosser gains over 7,000 votes after discovery of error
The San Francisco Examiner -- 04/07/11
In a news conference moments ago, Waukesha County Clerk Kathy Nickolaus confirmed reports that she had inadvertently failed to send votes from the entire city of Brookfield to the Associated Press for tabulation. The error, she said, came when a spreadsheet that she was importing into a Microsoft Access database didn't properly save the vote totals.
“This is not a case of extra votes or extra ballots being found,” Nickolaus said. “This is human error, which I apologize for.”
The city of Brookfield cast 14,315 votes on Tuesday, with Prosser winning 10,859 to 3,456. There were also two smaller typing errors in the county, Nickolaus said.
I've worked with Access for about 15 years, and am considered an Access Expert.
I've been racking my brain, trying to imagine how this could have happened as described, and I'm at a loss ... ???
What Clerk Kathy Nickolaus describes, is very curious ...
In my experience, when an Excel Import into an Access table "fails", it is usually on specific Excel rows that have some sort of data type problem (text was entered in a Number Field, etc); Those errors are saved off into a separate Errors Table, for you to resolve. (This could be the "human error" -- failing to notice the Errors Table ?)
But how this could happen on 10,000+ records is incredible
-- presumably the Excel spreadsheet you are importing would be "standardized and clean records", since it was following certain protocols, and data checks in the input side of the system. (The records of people voting in Brookfield.)
If such safe guards were in place ...
even the occasional "data type" import errors in a system like this, would be extremely rare.
Also such Mass failure like this, would be very slow and usually be very noticeable, if you're paying attention, at all. (10's of minutes)
Data can get "corrupted" in transmission, but that was not mentioned as the cause, nor as the fix. (Did Brookfield wire in their results, or was the xls sent by email, or what?)
As I said, it is very curious, to my Access data-crunching mind.
Anyone else out there, got a possible scenario, or story, where Access could (or did fail) on a mass scale like this? On supposedly clean data.
In my 15 years of experience -- Access Databases, have been very reliable, and rarely, rarely, ever fails to save a "valid" transaction. By default Access tries to save everything (even if you forget to click save) and it usually does. Like clockwork.